{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 1.0, "eval_steps": 500, "global_step": 5000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 296.6875, "completions/mean_terminated_length": 296.6875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.0002, "grad_norm": 0.08098223060369492, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 8555.0, "reward": 0.2841421067714691, "reward_std": 0.5165173411369324, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13078124821186066, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 409.0, "completions/mean_terminated_length": 409.0, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.0004, "grad_norm": 0.047587815672159195, "learning_rate": 4e-09, "loss": -0.0, "num_tokens": 19115.0, "reward": 0.4124077260494232, "reward_std": 0.3768385350704193, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07481250166893005, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9925000071525574, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 806.4375, "completions/mean_terminated_length": 775.357177734375, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.0006, "grad_norm": 4.615644044747569e-08, "learning_rate": 8e-09, "loss": -0.0, "num_tokens": 35842.0, "reward": 5.249999617262802e-07, "reward_std": 5.416025601334695e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.4000000059604645, "step": 3 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 318.5625, "completions/mean_terminated_length": 318.5625, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.0008, "grad_norm": 0.03698469698429108, "learning_rate": 1.2e-08, "loss": 0.0, "num_tokens": 45323.0, "reward": 0.43073225021362305, "reward_std": 0.21433010697364807, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04896249622106552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 409.875, "completions/mean_terminated_length": 409.875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.001, "grad_norm": 0.05333085358142853, "learning_rate": 1.6e-08, "loss": -0.0, "num_tokens": 55769.0, "reward": 0.29225093126296997, "reward_std": 0.37121957540512085, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1469999998807907, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.78125, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 430.1875, "completions/mean_terminated_length": 390.6000061035156, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.0012, "grad_norm": 0.047228485345840454, "learning_rate": 2e-08, "loss": -0.0, "num_tokens": 66796.0, "reward": 0.4782792031764984, "reward_std": 0.33270373940467834, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2065562605857849, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7668749690055847, "step": 6 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 799.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 551.3125, "completions/mean_terminated_length": 551.3125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.0014, "grad_norm": 0.051436908543109894, "learning_rate": 2.4e-08, "loss": 0.0, "num_tokens": 80273.0, "reward": 0.24017572402954102, "reward_std": 0.3347819745540619, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10535000264644623, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.6000000238418579, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 701.0625, "completions/mean_terminated_length": 679.5333862304688, "completions/min_length": 493.0, "completions/min_terminated_length": 493.0, "epoch": 0.0016, "grad_norm": 4.109421425368964e-08, "learning_rate": 2.8e-08, "loss": -0.0, "num_tokens": 95346.0, "reward": 8.62499916820525e-07, "reward_std": 4.627814007562847e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.675000011920929, "step": 8 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 365.3125, "completions/mean_terminated_length": 365.3125, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.0018, "grad_norm": 0.08075613528490067, "learning_rate": 3.2e-08, "loss": -0.0, "num_tokens": 105319.0, "reward": 0.29023557901382446, "reward_std": 0.5194936394691467, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14296874403953552, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.903124988079071, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 483.0625, "completions/mean_terminated_length": 483.0625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.002, "grad_norm": 0.041909489780664444, "learning_rate": 3.6e-08, "loss": -0.0, "num_tokens": 116984.0, "reward": 0.199454203248024, "reward_std": 0.4136039912700653, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08640624582767487, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 533.5, "completions/mean_terminated_length": 533.5, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.0022, "grad_norm": 0.04864685982465744, "learning_rate": 4e-08, "loss": 0.0, "num_tokens": 129216.0, "reward": 0.5806262493133545, "reward_std": 0.5102455019950867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22374999523162842, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 555.9375, "completions/mean_terminated_length": 555.9375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.0024, "grad_norm": 0.04805196821689606, "learning_rate": 4.4e-08, "loss": 0.0, "num_tokens": 141983.0, "reward": 0.39015746116638184, "reward_std": 0.5044029355049133, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15531249344348907, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9343750476837158, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 614.75, "completions/mean_terminated_length": 587.4666748046875, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.0026, "grad_norm": 0.01870528981089592, "learning_rate": 4.8e-08, "loss": 0.0, "num_tokens": 156347.0, "reward": 0.06616979837417603, "reward_std": 0.18094612658023834, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.007337499875575304, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.7325000166893005, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 425.8125, "completions/mean_terminated_length": 425.8125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0028, "grad_norm": 0.07845951616764069, "learning_rate": 5.1999999999999996e-08, "loss": -0.0, "num_tokens": 167848.0, "reward": 0.4808606803417206, "reward_std": 0.5627993941307068, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21171875298023224, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 323.0625, "completions/mean_terminated_length": 323.0625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.003, "grad_norm": 0.085394948720932, "learning_rate": 5.6e-08, "loss": 0.0, "num_tokens": 176713.0, "reward": 0.4165577292442322, "reward_std": 0.49284353852272034, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14561250805854797, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 382.0, "completions/max_terminated_length": 382.0, "completions/mean_length": 281.5625, "completions/mean_terminated_length": 281.5625, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.0032, "grad_norm": 0.06423453986644745, "learning_rate": 6e-08, "loss": -0.0, "num_tokens": 184738.0, "reward": 1.3780540227890015, "reward_std": 0.3313758969306946, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8811062574386597, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 16 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 407.8125, "completions/mean_terminated_length": 407.8125, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.0034, "grad_norm": 0.052097294479608536, "learning_rate": 6.4e-08, "loss": -0.0, "num_tokens": 195679.0, "reward": 0.5272666215896606, "reward_std": 0.44754868745803833, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1795312464237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 17 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 667.9375, "completions/mean_terminated_length": 585.7692260742188, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.0036, "grad_norm": 0.03273668512701988, "learning_rate": 6.8e-08, "loss": 0.0, "num_tokens": 211662.0, "reward": 0.15742263197898865, "reward_std": 0.29146426916122437, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06484374403953552, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.4999999701976776, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 564.125, "completions/mean_terminated_length": 564.125, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.0038, "grad_norm": 0.024335326626896858, "learning_rate": 7.2e-08, "loss": -0.0, "num_tokens": 227888.0, "reward": 0.19973865151405334, "reward_std": 0.26651960611343384, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02447500079870224, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 675.5, "completions/mean_terminated_length": 652.2667236328125, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.004, "grad_norm": 0.05805324390530586, "learning_rate": 7.599999999999999e-08, "loss": 0.0, "num_tokens": 243144.0, "reward": 0.21031305193901062, "reward_std": 0.3847746253013611, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1706250011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.4781250059604645, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 264.9375, "completions/mean_terminated_length": 264.9375, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 0.0042, "grad_norm": 0.14386257529258728, "learning_rate": 8e-08, "loss": -0.0, "num_tokens": 251623.0, "reward": 0.8745982646942139, "reward_std": 0.7412574887275696, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5616937279701233, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 21 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 374.0, "completions/mean_terminated_length": 374.0, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.0044, "grad_norm": 0.06871393322944641, "learning_rate": 8.4e-08, "loss": -0.0, "num_tokens": 261543.0, "reward": 0.5350011587142944, "reward_std": 0.49061721563339233, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19499997794628143, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8056249618530273, "step": 22 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 520.625, "completions/mean_terminated_length": 520.625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.0046, "grad_norm": 0.060024622827768326, "learning_rate": 8.8e-08, "loss": 0.0, "num_tokens": 273713.0, "reward": 0.21874484419822693, "reward_std": 0.5153716802597046, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12498749792575836, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 23 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 453.0, "completions/mean_terminated_length": 453.0, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.0048, "grad_norm": 0.09471607953310013, "learning_rate": 9.199999999999999e-08, "loss": -0.0, "num_tokens": 285025.0, "reward": 0.6559978723526001, "reward_std": 0.7682250738143921, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4369937479496002, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.84375, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 487.75, "completions/mean_terminated_length": 487.75, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.005, "grad_norm": 0.05350782349705696, "learning_rate": 9.6e-08, "loss": 0.0, "num_tokens": 297405.0, "reward": 0.19351713359355927, "reward_std": 0.40819352865219116, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07453124970197678, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.956250011920929, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 272.375, "completions/mean_terminated_length": 272.375, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.0052, "grad_norm": 0.08026593178510666, "learning_rate": 1e-07, "loss": 0.0, "num_tokens": 305939.0, "reward": 0.4148077368736267, "reward_std": 0.4925450086593628, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14211249351501465, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 26 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 311.0625, "completions/mean_terminated_length": 311.0625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.0054, "grad_norm": 0.11001398414373398, "learning_rate": 1.0399999999999999e-07, "loss": 0.0, "num_tokens": 314644.0, "reward": 0.628494918346405, "reward_std": 0.717992901802063, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38198748230934143, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.918749988079071, "step": 27 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 640.0625, "completions/mean_terminated_length": 614.4666748046875, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.0056, "grad_norm": 0.05613888427615166, "learning_rate": 1.0799999999999999e-07, "loss": 0.0, "num_tokens": 328693.0, "reward": 0.29804787039756775, "reward_std": 0.5206626057624817, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15859374403953552, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.690625011920929, "step": 28 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 461.75, "completions/mean_terminated_length": 461.75, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.0058, "grad_norm": 0.05268101394176483, "learning_rate": 1.12e-07, "loss": -0.0, "num_tokens": 339793.0, "reward": 0.20953227579593658, "reward_std": 0.42581894993782043, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10656249523162842, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.783750057220459, "step": 29 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 483.5, "completions/mean_terminated_length": 483.5, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.006, "grad_norm": 0.10266780853271484, "learning_rate": 1.16e-07, "loss": -0.0, "num_tokens": 351865.0, "reward": 0.5621699690818787, "reward_std": 0.6962147355079651, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37433749437332153, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.890625, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 653.5, "completions/mean_terminated_length": 600.5714721679688, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.0062, "grad_norm": 0.04812600463628769, "learning_rate": 1.2e-07, "loss": 0.0, "num_tokens": 368049.0, "reward": 0.23086017370224, "reward_std": 0.4161747992038727, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08671875298023224, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 31 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 796.875, "completions/mean_terminated_length": 744.4615478515625, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.0064, "grad_norm": 0.0131954625248909, "learning_rate": 1.24e-07, "loss": -0.0, "num_tokens": 384879.0, "reward": 0.06554816663265228, "reward_std": 0.17932923138141632, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.7312500476837158, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 490.9375, "completions/mean_terminated_length": 490.9375, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.0066, "grad_norm": 5.277951231619227e-08, "learning_rate": 1.28e-07, "loss": -0.0, "num_tokens": 396766.0, "reward": 9.706249102237052e-07, "reward_std": 5.169715677766362e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.7831249833106995, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 302.0, "completions/max_terminated_length": 302.0, "completions/mean_length": 186.0625, "completions/mean_terminated_length": 186.0625, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.0068, "grad_norm": 0.1526484340429306, "learning_rate": 1.32e-07, "loss": -0.0, "num_tokens": 403343.0, "reward": 0.9687390327453613, "reward_std": 0.7180604338645935, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6249749660491943, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9962500333786011, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 398.8125, "completions/mean_terminated_length": 398.8125, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.007, "grad_norm": 0.09476573765277863, "learning_rate": 1.36e-07, "loss": 0.0, "num_tokens": 414044.0, "reward": 0.6278013586997986, "reward_std": 0.7180137038230896, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3806000053882599, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 439.8125, "completions/mean_terminated_length": 439.8125, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0072, "grad_norm": 0.054635241627693176, "learning_rate": 1.4e-07, "loss": 0.0, "num_tokens": 425753.0, "reward": 0.7398197650909424, "reward_std": 0.567927896976471, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3546375036239624, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8974999785423279, "step": 36 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 398.8125, "completions/mean_terminated_length": 398.8125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.0074, "grad_norm": 0.03710181638598442, "learning_rate": 1.44e-07, "loss": 0.0, "num_tokens": 436278.0, "reward": 0.13703256845474243, "reward_std": 0.24574193358421326, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.024062499403953552, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9375, "step": 37 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 979.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 512.5, "completions/mean_terminated_length": 512.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0076, "grad_norm": 0.04099564999341965, "learning_rate": 1.4799999999999998e-07, "loss": -0.0, "num_tokens": 448206.0, "reward": 0.2752353250980377, "reward_std": 0.33826690912246704, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11296874284744263, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.828125, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 559.0, "completions/mean_terminated_length": 559.0, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.0078, "grad_norm": 0.05250097066164017, "learning_rate": 1.5199999999999998e-07, "loss": -0.0, "num_tokens": 460670.0, "reward": 0.45453232526779175, "reward_std": 0.39937448501586914, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2840625047683716, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.768750011920929, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 444.5625, "completions/mean_terminated_length": 444.5625, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.008, "grad_norm": 0.03609979525208473, "learning_rate": 1.56e-07, "loss": 0.0, "num_tokens": 471543.0, "reward": 0.13703233003616333, "reward_std": 0.2457420825958252, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.024062499403953552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.953125, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 485.0625, "completions/mean_terminated_length": 485.0625, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0082, "grad_norm": 0.059017959982156754, "learning_rate": 1.6e-07, "loss": 0.0, "num_tokens": 483256.0, "reward": 0.5210950374603271, "reward_std": 0.5484092831611633, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22968749701976776, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 41 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 938.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 630.5625, "completions/mean_terminated_length": 630.5625, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.0084, "grad_norm": 0.03726672753691673, "learning_rate": 1.64e-07, "loss": -0.0, "num_tokens": 497905.0, "reward": 0.16507907211780548, "reward_std": 0.31170397996902466, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08015625178813934, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.828125, "step": 42 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 436.75, "completions/mean_terminated_length": 436.75, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.0086, "grad_norm": 0.03551169112324715, "learning_rate": 1.68e-07, "loss": 0.0, "num_tokens": 508893.0, "reward": 0.2855362296104431, "reward_std": 0.2603638470172882, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.008568749763071537, "rewards/confidence_one_or_zero": 0.8125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9950000047683716, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 349.5, "completions/mean_terminated_length": 349.5, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.0088, "grad_norm": 0.09479932487010956, "learning_rate": 1.7199999999999998e-07, "loss": -0.0, "num_tokens": 519541.0, "reward": 0.8774104118347168, "reward_std": 0.7395034432411194, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5673187971115112, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 44 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 424.3125, "completions/mean_terminated_length": 424.3125, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.009, "grad_norm": 0.005391210783272982, "learning_rate": 1.76e-07, "loss": -0.0, "num_tokens": 530186.0, "reward": 0.5508604049682617, "reward_std": 0.03821711987257004, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10171874612569809, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 763.75, "completions/mean_terminated_length": 726.5714721679688, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.0092, "grad_norm": 5.5150554345573255e-08, "learning_rate": 1.8e-07, "loss": 0.0, "num_tokens": 546214.0, "reward": 1.0987499763359665e-06, "reward_std": 6.25266636689048e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.7862499952316284, "step": 46 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 531.9375, "completions/mean_terminated_length": 531.9375, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.0094, "grad_norm": 0.03360491618514061, "learning_rate": 1.8399999999999998e-07, "loss": 0.0, "num_tokens": 558741.0, "reward": 0.23217622935771942, "reward_std": 0.272300660610199, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.026850000023841858, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 511.4375, "completions/mean_terminated_length": 511.4375, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0096, "grad_norm": 0.07670846581459045, "learning_rate": 1.88e-07, "loss": 0.0, "num_tokens": 570604.0, "reward": 0.38403865694999695, "reward_std": 0.593782901763916, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20557500422000885, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 594.0625, "completions/mean_terminated_length": 594.0625, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.0098, "grad_norm": 0.05169695243239403, "learning_rate": 1.92e-07, "loss": -0.0, "num_tokens": 584349.0, "reward": 0.4226450026035309, "reward_std": 0.4942355751991272, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1577875018119812, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 49 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 544.625, "completions/mean_terminated_length": 544.625, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.01, "grad_norm": 0.04680353403091431, "learning_rate": 1.96e-07, "loss": 0.0, "num_tokens": 596999.0, "reward": 0.1678917109966278, "reward_std": 0.40860599279403687, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08578124642372131, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 542.6875, "completions/mean_terminated_length": 542.6875, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.0102, "grad_norm": 0.06311339139938354, "learning_rate": 2e-07, "loss": 0.0, "num_tokens": 609778.0, "reward": 0.1965635120868683, "reward_std": 0.41080746054649353, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08062499761581421, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.762499988079071, "step": 51 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 294.125, "completions/mean_terminated_length": 294.125, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0104, "grad_norm": 0.0767366886138916, "learning_rate": 2.0399999999999997e-07, "loss": -0.0, "num_tokens": 618372.0, "reward": 0.5793482661247253, "reward_std": 0.5107517242431641, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22119373083114624, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 52 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 496.75, "completions/mean_terminated_length": 496.75, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.0106, "grad_norm": 0.03687542676925659, "learning_rate": 2.0799999999999998e-07, "loss": -0.0, "num_tokens": 630160.0, "reward": 0.1623326539993286, "reward_std": 0.40249451994895935, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07466249912977219, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 53 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 588.9375, "completions/mean_terminated_length": 588.9375, "completions/min_length": 457.0, "completions/min_terminated_length": 457.0, "epoch": 0.0108, "grad_norm": 3.817957505702907e-08, "learning_rate": 2.12e-07, "loss": 0.0, "num_tokens": 643423.0, "reward": 1.1375000212865416e-06, "reward_std": 4.3327435150786187e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.8875000476837158, "step": 54 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 330.0, "completions/mean_terminated_length": 330.0, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.011, "grad_norm": 0.055453937500715256, "learning_rate": 2.1599999999999998e-07, "loss": -0.0, "num_tokens": 652767.0, "reward": 0.39567920565605164, "reward_std": 0.39483457803726196, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10385624319314957, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 241.5, "completions/mean_terminated_length": 241.5, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.0112, "grad_norm": 0.12136227637529373, "learning_rate": 2.1999999999999998e-07, "loss": 0.0, "num_tokens": 660535.0, "reward": 0.9307044744491577, "reward_std": 0.5336164236068726, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6114062070846558, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 335.5625, "completions/mean_terminated_length": 335.5625, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.0114, "grad_norm": 0.09655342251062393, "learning_rate": 2.24e-07, "loss": 0.0, "num_tokens": 669808.0, "reward": 0.5714478492736816, "reward_std": 0.6791788339614868, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3303937315940857, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 57 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 549.625, "completions/mean_terminated_length": 549.625, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.0116, "grad_norm": 0.05402513965964317, "learning_rate": 2.28e-07, "loss": -0.0, "num_tokens": 682490.0, "reward": 0.3237389326095581, "reward_std": 0.6074484586715698, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.209975004196167, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 58 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 942.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 646.25, "completions/mean_terminated_length": 646.25, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.0118, "grad_norm": 0.024387778714299202, "learning_rate": 2.32e-07, "loss": 0.0, "num_tokens": 696974.0, "reward": 0.0974198430776596, "reward_std": 0.20965522527694702, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.007337499875575304, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.7725000381469727, "step": 59 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 444.3125, "completions/mean_terminated_length": 444.3125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.012, "grad_norm": 0.04218657314777374, "learning_rate": 2.3599999999999997e-07, "loss": 0.0, "num_tokens": 707811.0, "reward": 0.3458603322505951, "reward_std": 0.27691933512687683, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06671874970197678, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9393750429153442, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 439.375, "completions/mean_terminated_length": 439.375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.0122, "grad_norm": 0.03503440320491791, "learning_rate": 2.4e-07, "loss": -0.0, "num_tokens": 719369.0, "reward": 0.23522308468818665, "reward_std": 0.27574801445007324, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03294374793767929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 61 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 564.1875, "completions/mean_terminated_length": 533.5333862304688, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.0124, "grad_norm": 0.027055826038122177, "learning_rate": 2.4399999999999996e-07, "loss": 0.0, "num_tokens": 732268.0, "reward": 0.10289178788661957, "reward_std": 0.22120808064937592, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.783750057220459, "step": 62 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 549.75, "completions/mean_terminated_length": 482.0000305175781, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.0126, "grad_norm": 0.039464011788368225, "learning_rate": 2.48e-07, "loss": 0.0, "num_tokens": 745928.0, "reward": 0.20343813300132751, "reward_std": 0.3690200448036194, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15687499940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.44062501192092896, "step": 63 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 316.0, "completions/max_terminated_length": 316.0, "completions/mean_length": 243.8125, "completions/mean_terminated_length": 243.8125, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.0128, "grad_norm": 0.10453317314386368, "learning_rate": 2.52e-07, "loss": -0.0, "num_tokens": 753605.0, "reward": 0.35585731267929077, "reward_std": 0.5114989876747131, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14921249449253082, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 64 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 613.5625, "completions/mean_terminated_length": 613.5625, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.013, "grad_norm": 0.07198241353034973, "learning_rate": 2.56e-07, "loss": -0.0, "num_tokens": 767390.0, "reward": 0.6590637564659119, "reward_std": 0.6995270252227783, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38062500953674316, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 65 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 322.375, "completions/mean_terminated_length": 322.375, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.0132, "grad_norm": 0.03976324573159218, "learning_rate": 2.6e-07, "loss": 0.0, "num_tokens": 776356.0, "reward": 0.1653919667005539, "reward_std": 0.2537340819835663, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 66 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 442.625, "completions/mean_terminated_length": 442.625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.0134, "grad_norm": 0.06899787485599518, "learning_rate": 2.64e-07, "loss": 0.0, "num_tokens": 787662.0, "reward": 0.3845195472240448, "reward_std": 0.49682509899139404, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.26903748512268066, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.5518749952316284, "step": 67 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 349.8125, "completions/mean_terminated_length": 349.8125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.0136, "grad_norm": 0.0802777037024498, "learning_rate": 2.68e-07, "loss": -0.0, "num_tokens": 797147.0, "reward": 0.2560917139053345, "reward_std": 0.5197939872741699, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13718125224113464, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 68 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 438.1875, "completions/mean_terminated_length": 438.1875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.0138, "grad_norm": 0.030244166031479836, "learning_rate": 2.72e-07, "loss": 0.0, "num_tokens": 807838.0, "reward": 0.1310950517654419, "reward_std": 0.23484528064727783, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9231250286102295, "step": 69 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 317.5625, "completions/mean_terminated_length": 317.5625, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.014, "grad_norm": 0.0944381058216095, "learning_rate": 2.7600000000000004e-07, "loss": 0.0, "num_tokens": 817383.0, "reward": 0.521160364151001, "reward_std": 0.5479705333709717, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22981873154640198, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 70 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 719.6875, "completions/mean_terminated_length": 676.2142944335938, "completions/min_length": 519.0, "completions/min_terminated_length": 519.0, "epoch": 0.0142, "grad_norm": 0.01613769307732582, "learning_rate": 2.8e-07, "loss": 0.0, "num_tokens": 833394.0, "reward": 0.0685947835445404, "reward_std": 0.18743376433849335, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.846875011920929, "step": 71 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 353.5625, "completions/mean_terminated_length": 353.5625, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.0144, "grad_norm": 0.10841629654169083, "learning_rate": 2.8399999999999995e-07, "loss": 0.0, "num_tokens": 844235.0, "reward": 0.8153108358383179, "reward_std": 0.7259045243263245, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5056187510490417, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 72 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 463.625, "completions/mean_terminated_length": 463.625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.0146, "grad_norm": 0.06379537284374237, "learning_rate": 2.88e-07, "loss": 0.0, "num_tokens": 855701.0, "reward": 0.2964072823524475, "reward_std": 0.41602709889411926, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09281250089406967, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 73 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 566.3125, "completions/mean_terminated_length": 566.3125, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.0148, "grad_norm": 0.0370805524289608, "learning_rate": 2.9199999999999997e-07, "loss": 0.0, "num_tokens": 869754.0, "reward": 0.2797665297985077, "reward_std": 0.3362070918083191, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12203125655651093, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.5874999761581421, "step": 74 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 758.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 542.0625, "completions/mean_terminated_length": 542.0625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.015, "grad_norm": 0.03968357294797897, "learning_rate": 2.9599999999999995e-07, "loss": -0.0, "num_tokens": 882363.0, "reward": 0.10117287933826447, "reward_std": 0.28338760137557983, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07734374701976776, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.8793749809265137, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 642.0625, "completions/mean_terminated_length": 616.6000366210938, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.0152, "grad_norm": 0.026430975645780563, "learning_rate": 3e-07, "loss": 0.0, "num_tokens": 896268.0, "reward": 0.19968855381011963, "reward_std": 0.2666465938091278, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02437499910593033, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.734375, "step": 76 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 884.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 532.1875, "completions/mean_terminated_length": 532.1875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.0154, "grad_norm": 0.06590546667575836, "learning_rate": 3.0399999999999997e-07, "loss": -0.0, "num_tokens": 908623.0, "reward": 0.2847635746002197, "reward_std": 0.5168006420135498, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13202500343322754, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 77 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 289.125, "completions/mean_terminated_length": 289.125, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.0156, "grad_norm": 0.11769720166921616, "learning_rate": 3.08e-07, "loss": -0.0, "num_tokens": 917665.0, "reward": 0.6056228876113892, "reward_std": 0.6618617177009583, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33624374866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9543749690055847, "step": 78 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 895.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 598.625, "completions/mean_terminated_length": 598.625, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.0158, "grad_norm": 0.051605500280857086, "learning_rate": 3.12e-07, "loss": 0.0, "num_tokens": 931499.0, "reward": 0.2745320498943329, "reward_std": 0.46587368845939636, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17406249046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.7781250476837158, "step": 79 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 438.375, "completions/mean_terminated_length": 438.375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.016, "grad_norm": 0.08335080742835999, "learning_rate": 3.1599999999999997e-07, "loss": 0.0, "num_tokens": 942385.0, "reward": 0.44023552536964417, "reward_std": 0.6549926400184631, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2554687559604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 80 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 430.125, "completions/mean_terminated_length": 390.5333557128906, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.0162, "grad_norm": 0.05784187465906143, "learning_rate": 3.2e-07, "loss": -0.0, "num_tokens": 953299.0, "reward": 0.26515713334083557, "reward_std": 0.4184861183166504, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09281250089406967, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 81 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 372.75, "completions/mean_terminated_length": 372.75, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.0164, "grad_norm": 0.08087868988513947, "learning_rate": 3.24e-07, "loss": 0.0, "num_tokens": 963663.0, "reward": 0.5088949203491211, "reward_std": 0.632348358631134, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26778751611709595, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 82 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 375.5625, "completions/mean_terminated_length": 375.5625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.0166, "grad_norm": 0.09731743484735489, "learning_rate": 3.28e-07, "loss": 0.0, "num_tokens": 973432.0, "reward": 0.7494542002677917, "reward_std": 0.774031937122345, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.4989062547683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 83 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 653.8125, "completions/mean_terminated_length": 653.8125, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.0168, "grad_norm": 0.01268667634576559, "learning_rate": 3.32e-07, "loss": -0.0, "num_tokens": 988405.0, "reward": 0.03125102445483208, "reward_std": 0.1250002682209015, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.7112500071525574, "step": 84 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 406.75, "completions/mean_terminated_length": 406.75, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.017, "grad_norm": 0.04361812025308609, "learning_rate": 3.36e-07, "loss": -0.0, "num_tokens": 999153.0, "reward": 0.21843892335891724, "reward_std": 0.3137500584125519, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06187500059604645, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.908750057220459, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 382.9375, "completions/mean_terminated_length": 382.9375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.0172, "grad_norm": 0.09060061722993851, "learning_rate": 3.4000000000000003e-07, "loss": 0.0, "num_tokens": 1008960.0, "reward": 0.6033478379249573, "reward_std": 0.6623716354370117, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33169376850128174, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 86 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 813.0625, "completions/mean_terminated_length": 764.3846435546875, "completions/min_length": 621.0, "completions/min_terminated_length": 621.0, "epoch": 0.0174, "grad_norm": 0.02790345996618271, "learning_rate": 3.4399999999999996e-07, "loss": 0.0, "num_tokens": 1026449.0, "reward": 0.16203199326992035, "reward_std": 0.30791911482810974, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07406249642372131, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5437500476837158, "step": 87 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 354.6875, "completions/mean_terminated_length": 354.6875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.0176, "grad_norm": 0.06653674691915512, "learning_rate": 3.4799999999999994e-07, "loss": 0.0, "num_tokens": 1036476.0, "reward": 0.4690476655960083, "reward_std": 0.4504992365837097, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.25059375166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7231249809265137, "step": 88 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 440.0625, "completions/mean_terminated_length": 440.0625, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.0178, "grad_norm": 0.015943417325615883, "learning_rate": 3.52e-07, "loss": 0.0, "num_tokens": 1047213.0, "reward": 0.03429817780852318, "reward_std": 0.1371873915195465, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 89 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 680.75, "completions/mean_terminated_length": 631.7142944335938, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.018, "grad_norm": 0.0328657366335392, "learning_rate": 3.5599999999999996e-07, "loss": 0.0, "num_tokens": 1062313.0, "reward": 0.17648503184318542, "reward_std": 0.3800133168697357, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1654687523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.65625, "step": 90 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 882.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 505.125, "completions/mean_terminated_length": 505.125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.0182, "grad_norm": 0.05307362228631973, "learning_rate": 3.6e-07, "loss": 0.0, "num_tokens": 1074235.0, "reward": 0.5252888202667236, "reward_std": 0.48488855361938477, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23807500302791595, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.908750057220459, "step": 91 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 330.1875, "completions/mean_terminated_length": 330.1875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0184, "grad_norm": 0.10249677300453186, "learning_rate": 3.64e-07, "loss": -0.0, "num_tokens": 1084254.0, "reward": 0.37476682662963867, "reward_std": 0.6704011559486389, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.24953123927116394, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8512499928474426, "step": 92 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 732.6875, "completions/mean_terminated_length": 732.6875, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.0186, "grad_norm": 0.0318664014339447, "learning_rate": 3.6799999999999996e-07, "loss": 0.0, "num_tokens": 1100937.0, "reward": 0.21648527681827545, "reward_std": 0.43418172001838684, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12046875059604645, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5268750190734863, "step": 93 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 517.0625, "completions/mean_terminated_length": 517.0625, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.0188, "grad_norm": 0.053945332765579224, "learning_rate": 3.72e-07, "loss": 0.0, "num_tokens": 1113274.0, "reward": 0.15625160932540894, "reward_std": 0.39660027623176575, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0625, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 94 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 455.3125, "completions/mean_terminated_length": 455.3125, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.019, "grad_norm": 0.0777345523238182, "learning_rate": 3.76e-07, "loss": -0.0, "num_tokens": 1124319.0, "reward": 0.5621885657310486, "reward_std": 0.7495835423469543, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.37437498569488525, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 95 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 346.375, "completions/mean_terminated_length": 346.375, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.0192, "grad_norm": 0.03509316220879555, "learning_rate": 3.7999999999999996e-07, "loss": 0.0, "num_tokens": 1133829.0, "reward": 0.4771103858947754, "reward_std": 0.1866389811038971, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.953125, "step": 96 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 397.9375, "completions/mean_terminated_length": 397.9375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.0194, "grad_norm": 0.09883241355419159, "learning_rate": 3.84e-07, "loss": 0.0, "num_tokens": 1144756.0, "reward": 0.4683603048324585, "reward_std": 0.7174719572067261, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31171876192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.949999988079071, "step": 97 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 673.0, "completions/mean_terminated_length": 592.0, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.0196, "grad_norm": 0.018823925405740738, "learning_rate": 3.88e-07, "loss": 0.0, "num_tokens": 1161412.0, "reward": 0.06554779410362244, "reward_std": 0.17932938039302826, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 98 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 475.0, "completions/max_terminated_length": 475.0, "completions/mean_length": 357.5625, "completions/mean_terminated_length": 357.5625, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0198, "grad_norm": 0.09500116854906082, "learning_rate": 3.92e-07, "loss": -0.0, "num_tokens": 1171405.0, "reward": 0.5353635549545288, "reward_std": 0.6941308379173279, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3207250237464905, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 99 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 326.75, "completions/mean_terminated_length": 326.75, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.02, "grad_norm": 0.11517077684402466, "learning_rate": 3.96e-07, "loss": -0.0, "num_tokens": 1180441.0, "reward": 0.7893602848052979, "reward_std": 0.6765143275260925, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45371872186660767, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8887499570846558, "step": 100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 611.0625, "completions/mean_terminated_length": 611.0625, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.0202, "grad_norm": 0.07177139073610306, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 1195770.0, "reward": 0.2663291096687317, "reward_std": 0.5268847346305847, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1576562523841858, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 553.9375, "completions/mean_terminated_length": 522.6000366210938, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.0204, "grad_norm": 0.04729292169213295, "learning_rate": 4.04e-07, "loss": 0.0, "num_tokens": 1208969.0, "reward": 0.2808602452278137, "reward_std": 0.4025823175907135, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18671876192092896, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.4906250238418579, "step": 102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 287.125, "completions/mean_terminated_length": 287.125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.0206, "grad_norm": 0.11902254819869995, "learning_rate": 4.0799999999999995e-07, "loss": 0.0, "num_tokens": 1217851.0, "reward": 0.3786671757698059, "reward_std": 0.592555046081543, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1948312371969223, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9906250238418579, "step": 103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 340.25, "completions/mean_terminated_length": 340.25, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.0208, "grad_norm": 0.12371638417243958, "learning_rate": 4.12e-07, "loss": 0.0, "num_tokens": 1226863.0, "reward": 0.5660855770111084, "reward_std": 0.6797011494636536, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3196687400341034, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9718750715255737, "step": 104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 352.6875, "completions/mean_terminated_length": 352.6875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.021, "grad_norm": 0.09116816520690918, "learning_rate": 4.1599999999999997e-07, "loss": 0.0, "num_tokens": 1236858.0, "reward": 0.7586603164672852, "reward_std": 0.7033853530883789, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4548187553882599, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 623.25, "completions/mean_terminated_length": 623.25, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.0212, "grad_norm": 0.04949285089969635, "learning_rate": 4.1999999999999995e-07, "loss": 0.0, "num_tokens": 1251118.0, "reward": 0.382954478263855, "reward_std": 0.5023974180221558, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14090624451637268, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 398.125, "completions/mean_terminated_length": 398.125, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.0214, "grad_norm": 0.10399240255355835, "learning_rate": 4.24e-07, "loss": -0.0, "num_tokens": 1261376.0, "reward": 0.6993760466575623, "reward_std": 0.6764371991157532, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3987500071525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 357.1875, "completions/mean_terminated_length": 357.1875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.0216, "grad_norm": 0.11609475314617157, "learning_rate": 4.2799999999999997e-07, "loss": -0.0, "num_tokens": 1270595.0, "reward": 0.8754479885101318, "reward_std": 0.7411313652992249, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.563393771648407, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 401.0625, "completions/mean_terminated_length": 401.0625, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0218, "grad_norm": 0.07339116930961609, "learning_rate": 4.3199999999999995e-07, "loss": -0.0, "num_tokens": 1281236.0, "reward": 0.5052104592323303, "reward_std": 0.632072389125824, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26041874289512634, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9043750166893005, "step": 109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 562.5, "completions/mean_terminated_length": 496.5714416503906, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.022, "grad_norm": 0.043381135910749435, "learning_rate": 4.36e-07, "loss": 0.0, "num_tokens": 1294524.0, "reward": 0.2339852750301361, "reward_std": 0.2744274437427521, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 350.875, "completions/mean_terminated_length": 350.875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.0222, "grad_norm": 0.062040649354457855, "learning_rate": 4.3999999999999997e-07, "loss": 0.0, "num_tokens": 1303994.0, "reward": 0.397488534450531, "reward_std": 0.39550718665122986, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10747499763965607, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 273.0, "completions/max_terminated_length": 273.0, "completions/mean_length": 182.375, "completions/mean_terminated_length": 182.375, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 0.0224, "grad_norm": 0.13398364186286926, "learning_rate": 4.44e-07, "loss": 0.0, "num_tokens": 1310528.0, "reward": 1.0724074840545654, "reward_std": 0.591610848903656, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6448124647140503, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 499.4375, "completions/mean_terminated_length": 499.4375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.0226, "grad_norm": 0.03705069050192833, "learning_rate": 4.48e-07, "loss": -0.0, "num_tokens": 1325671.0, "reward": 0.06218857318162918, "reward_std": 0.2487497329711914, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06187500059604645, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.5706250071525574, "step": 113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 552.0, "completions/mean_terminated_length": 552.0, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.0228, "grad_norm": 0.05325879529118538, "learning_rate": 4.5199999999999997e-07, "loss": -0.0, "num_tokens": 1339111.0, "reward": 0.21367251873016357, "reward_std": 0.3938092291355133, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17734375596046448, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.4768750071525574, "step": 114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 298.375, "completions/mean_terminated_length": 298.375, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.023, "grad_norm": 0.017881715670228004, "learning_rate": 4.56e-07, "loss": -0.0, "num_tokens": 1347725.0, "reward": 0.03429822623729706, "reward_std": 0.1371873915195465, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 466.0, "completions/max_terminated_length": 466.0, "completions/mean_length": 368.5625, "completions/mean_terminated_length": 368.5625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.0232, "grad_norm": 0.06430233269929886, "learning_rate": 4.6e-07, "loss": -0.0, "num_tokens": 1357446.0, "reward": 0.2559387683868408, "reward_std": 0.519402801990509, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1368750035762787, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 652.5, "completions/mean_terminated_length": 652.5, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.0234, "grad_norm": 0.050697069615125656, "learning_rate": 4.64e-07, "loss": 0.0, "num_tokens": 1372046.0, "reward": 0.31553229689598083, "reward_std": 0.6031815409660339, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19356249272823334, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7931250333786011, "step": 117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 386.0625, "completions/mean_terminated_length": 386.0625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.0236, "grad_norm": 0.07440226525068283, "learning_rate": 4.68e-07, "loss": -0.0, "num_tokens": 1382303.0, "reward": 0.6976197957992554, "reward_std": 0.6769724488258362, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3952375054359436, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 403.625, "completions/mean_terminated_length": 403.625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.0238, "grad_norm": 0.0342307947576046, "learning_rate": 4.7199999999999994e-07, "loss": -0.0, "num_tokens": 1392569.0, "reward": 0.102891705930233, "reward_std": 0.2212081253528595, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 372.125, "completions/mean_terminated_length": 372.125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.024, "grad_norm": 0.10766858607530594, "learning_rate": 4.76e-07, "loss": -0.0, "num_tokens": 1402443.0, "reward": 0.5323290824890137, "reward_std": 0.6941748261451721, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31465625762939453, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 285.4375, "completions/mean_terminated_length": 285.4375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.0242, "grad_norm": 0.10211831331253052, "learning_rate": 4.8e-07, "loss": -0.0, "num_tokens": 1411538.0, "reward": 0.6147511005401611, "reward_std": 0.5805013179779053, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2919999957084656, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 653.0625, "completions/mean_terminated_length": 628.3333740234375, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.0244, "grad_norm": 0.04420475661754608, "learning_rate": 4.839999999999999e-07, "loss": -0.0, "num_tokens": 1426387.0, "reward": 0.14836011826992035, "reward_std": 0.4210865795612335, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10921874642372131, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.565000057220459, "step": 122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 838.25, "completions/mean_terminated_length": 776.3333740234375, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.0246, "grad_norm": 0.039021607488393784, "learning_rate": 4.879999999999999e-07, "loss": -0.0, "num_tokens": 1444167.0, "reward": 0.16812554001808167, "reward_std": 0.36293184757232666, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14875000715255737, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.5406249761581421, "step": 123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1018.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 622.625, "completions/mean_terminated_length": 622.625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.0248, "grad_norm": 0.0340426079928875, "learning_rate": 4.92e-07, "loss": 0.0, "num_tokens": 1457905.0, "reward": 0.2186727374792099, "reward_std": 0.2955875098705292, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.062343746423721313, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8656250238418579, "step": 124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 698.625, "completions/mean_terminated_length": 676.933349609375, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.025, "grad_norm": 0.03701069951057434, "learning_rate": 4.96e-07, "loss": 0.0, "num_tokens": 1473227.0, "reward": 0.23711006343364716, "reward_std": 0.36462700366973877, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16171875596046448, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 475.875, "completions/mean_terminated_length": 475.875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.0252, "grad_norm": 0.07473543286323547, "learning_rate": 5e-07, "loss": -0.0, "num_tokens": 1484521.0, "reward": 0.4120166301727295, "reward_std": 0.6648141145706177, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.261531263589859, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 808.5625, "completions/mean_terminated_length": 777.7857666015625, "completions/min_length": 553.0, "completions/min_terminated_length": 553.0, "epoch": 0.0254, "grad_norm": 0.030593913048505783, "learning_rate": 5.04e-07, "loss": -0.0, "num_tokens": 1501746.0, "reward": 0.15898506343364716, "reward_std": 0.3546580374240875, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13046875596046448, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.5081249475479126, "step": 127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 735.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 413.0, "completions/mean_terminated_length": 413.0, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0256, "grad_norm": 0.05094014108181, "learning_rate": 5.079999999999999e-07, "loss": 0.0, "num_tokens": 1512018.0, "reward": 0.33202630281448364, "reward_std": 0.2662087380886078, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03904999792575836, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 536.25, "completions/mean_terminated_length": 536.25, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.0258, "grad_norm": 0.0766674280166626, "learning_rate": 5.12e-07, "loss": 0.0, "num_tokens": 1524886.0, "reward": 0.5028668642044067, "reward_std": 0.7069392800331116, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3182312250137329, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9106250405311584, "step": 129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 556.75, "completions/mean_terminated_length": 525.6000366210938, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.026, "grad_norm": 0.05938326567411423, "learning_rate": 5.16e-07, "loss": 0.0, "num_tokens": 1537330.0, "reward": 0.2214856892824173, "reward_std": 0.40835168957710266, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06796874850988388, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8062500357627869, "step": 130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 430.8125, "completions/mean_terminated_length": 430.8125, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0262, "grad_norm": 0.0666564553976059, "learning_rate": 5.2e-07, "loss": -0.0, "num_tokens": 1547951.0, "reward": 0.4472608268260956, "reward_std": 0.5751882195472717, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20701874792575836, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.984375, "step": 131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 430.8125, "completions/mean_terminated_length": 430.8125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0264, "grad_norm": 0.055563800036907196, "learning_rate": 5.24e-07, "loss": 0.0, "num_tokens": 1560796.0, "reward": 0.30390694737434387, "reward_std": 0.36234280467033386, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17031249403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.6906249523162842, "step": 132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 425.5625, "completions/mean_terminated_length": 425.5625, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0266, "grad_norm": 0.042575735598802567, "learning_rate": 5.28e-07, "loss": -0.0, "num_tokens": 1571797.0, "reward": 0.25953221321105957, "reward_std": 0.3060724437236786, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08156249672174454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 457.8125, "completions/mean_terminated_length": 457.8125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.0268, "grad_norm": 0.10171791166067123, "learning_rate": 5.32e-07, "loss": 0.0, "num_tokens": 1582754.0, "reward": 0.9575790762901306, "reward_std": 0.6504940390586853, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6026562452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 821.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 565.875, "completions/mean_terminated_length": 565.875, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.027, "grad_norm": 0.06400745362043381, "learning_rate": 5.36e-07, "loss": -0.0, "num_tokens": 1595904.0, "reward": 0.5497665405273438, "reward_std": 0.6200970411300659, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.41203123331069946, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.6499999761581421, "step": 135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 449.1875, "completions/mean_terminated_length": 410.86669921875, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.0272, "grad_norm": 0.0843084454536438, "learning_rate": 5.4e-07, "loss": 0.0, "num_tokens": 1607491.0, "reward": 0.5655478239059448, "reward_std": 0.6524590849876404, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38109374046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.734375, "step": 136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 735.75, "completions/mean_terminated_length": 694.5714721679688, "completions/min_length": 443.0, "completions/min_terminated_length": 443.0, "epoch": 0.0274, "grad_norm": 0.01874314807355404, "learning_rate": 5.44e-07, "loss": -0.0, "num_tokens": 1623167.0, "reward": 0.03429777920246124, "reward_std": 0.13718749582767487, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 304.25, "completions/mean_terminated_length": 304.25, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.0276, "grad_norm": 0.09558682888746262, "learning_rate": 5.48e-07, "loss": -0.0, "num_tokens": 1632083.0, "reward": 0.8188761472702026, "reward_std": 0.6515138745307922, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4502499997615814, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 280.0, "completions/max_terminated_length": 280.0, "completions/mean_length": 199.1875, "completions/mean_terminated_length": 199.1875, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.0278, "grad_norm": 0.15085892379283905, "learning_rate": 5.520000000000001e-07, "loss": 0.0, "num_tokens": 1638774.0, "reward": 1.0997610092163086, "reward_std": 0.5475200414657593, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6370187997817993, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9893749952316284, "step": 139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 441.875, "completions/mean_terminated_length": 441.875, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.028, "grad_norm": 0.04552271589636803, "learning_rate": 5.560000000000001e-07, "loss": 0.0, "num_tokens": 1649844.0, "reward": 0.39880090951919556, "reward_std": 0.39610251784324646, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11009999364614487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 440.25, "completions/mean_terminated_length": 401.3333435058594, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.0282, "grad_norm": 0.05709691345691681, "learning_rate": 5.6e-07, "loss": 0.0, "num_tokens": 1660856.0, "reward": 0.3289729058742523, "reward_std": 0.41184690594673157, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09544374793767929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 250.25, "completions/mean_terminated_length": 250.25, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.0284, "grad_norm": 0.11838651448488235, "learning_rate": 5.639999999999999e-07, "loss": -0.0, "num_tokens": 1668844.0, "reward": 0.8868259191513062, "reward_std": 0.6635146141052246, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5236499905586243, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9543749690055847, "step": 142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 694.4375, "completions/mean_terminated_length": 647.357177734375, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.0286, "grad_norm": 0.032341569662094116, "learning_rate": 5.679999999999999e-07, "loss": 0.0, "num_tokens": 1683955.0, "reward": 0.16311325132846832, "reward_std": 0.4027862548828125, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07622499763965607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.75062495470047, "step": 143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 559.8125, "completions/mean_terminated_length": 528.86669921875, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0288, "grad_norm": 0.040057819336652756, "learning_rate": 5.719999999999999e-07, "loss": 0.0, "num_tokens": 1699184.0, "reward": 0.3063920736312866, "reward_std": 0.3381233811378479, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11278124898672104, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9168750047683716, "step": 144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 535.4375, "completions/mean_terminated_length": 535.4375, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.029, "grad_norm": 0.03513384610414505, "learning_rate": 5.76e-07, "loss": 0.0, "num_tokens": 1711463.0, "reward": 0.13367316126823425, "reward_std": 0.30102962255477905, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07984375208616257, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.8550000190734863, "step": 145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 483.0, "completions/max_terminated_length": 483.0, "completions/mean_length": 368.375, "completions/mean_terminated_length": 368.375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.0292, "grad_norm": 0.08456844091415405, "learning_rate": 5.8e-07, "loss": -0.0, "num_tokens": 1721421.0, "reward": 0.39609473943710327, "reward_std": 0.5063376426696777, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16718749701976776, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.890625, "step": 146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 739.5625, "completions/mean_terminated_length": 698.9285888671875, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.0294, "grad_norm": 0.04048054292798042, "learning_rate": 5.839999999999999e-07, "loss": -0.0, "num_tokens": 1737302.0, "reward": 0.24867287278175354, "reward_std": 0.4807804524898529, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1848437488079071, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 247.8125, "completions/mean_terminated_length": 247.8125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.0296, "grad_norm": 0.14909808337688446, "learning_rate": 5.879999999999999e-07, "loss": -0.0, "num_tokens": 1745299.0, "reward": 1.1246012449264526, "reward_std": 0.6705821752548218, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7491999864578247, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 285.125, "completions/mean_terminated_length": 285.125, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.0298, "grad_norm": 0.13573239743709564, "learning_rate": 5.919999999999999e-07, "loss": -0.0, "num_tokens": 1754101.0, "reward": 0.5370978713035583, "reward_std": 0.6940144300460815, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3241937458515167, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 359.125, "completions/mean_terminated_length": 359.125, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.03, "grad_norm": 0.077030710875988, "learning_rate": 5.96e-07, "loss": 0.0, "num_tokens": 1764151.0, "reward": 0.3822481632232666, "reward_std": 0.5020091533660889, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13949374854564667, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 382.6875, "completions/mean_terminated_length": 382.6875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0302, "grad_norm": 0.06716588139533997, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 1774210.0, "reward": 0.2842206358909607, "reward_std": 0.4086473882198334, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06843750178813934, "rewards/confidence_one_or_zero": 0.875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9937499761581421, "step": 151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 859.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 585.375, "completions/mean_terminated_length": 585.375, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.0304, "grad_norm": 0.05125172808766365, "learning_rate": 6.04e-07, "loss": 0.0, "num_tokens": 1788328.0, "reward": 0.3841387927532196, "reward_std": 0.5028085112571716, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14327499270439148, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9706249833106995, "step": 152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 353.8125, "completions/mean_terminated_length": 353.8125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0306, "grad_norm": 0.04311644658446312, "learning_rate": 6.079999999999999e-07, "loss": 0.0, "num_tokens": 1797637.0, "reward": 0.3320261836051941, "reward_std": 0.2662088871002197, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03904999792575836, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9099999666213989, "step": 153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 297.8125, "completions/mean_terminated_length": 297.8125, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.0308, "grad_norm": 0.09323222190141678, "learning_rate": 6.119999999999999e-07, "loss": 0.0, "num_tokens": 1805762.0, "reward": 0.7128010988235474, "reward_std": 0.5023190975189209, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30059999227523804, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 689.625, "completions/mean_terminated_length": 689.625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.031, "grad_norm": 0.04638264700770378, "learning_rate": 6.16e-07, "loss": -0.0, "num_tokens": 1820700.0, "reward": 0.22172003984451294, "reward_std": 0.4480200409889221, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13093750178813934, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.6781250238418579, "step": 155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 554.5, "completions/mean_terminated_length": 554.5, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.0312, "grad_norm": 0.034764163196086884, "learning_rate": 6.2e-07, "loss": -0.0, "num_tokens": 1833684.0, "reward": 0.09367305040359497, "reward_std": 0.37468743324279785, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06234375014901161, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.861875057220459, "step": 156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 299.25, "completions/mean_terminated_length": 299.25, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.0314, "grad_norm": 0.09169171005487442, "learning_rate": 6.24e-07, "loss": 0.0, "num_tokens": 1842136.0, "reward": 0.4424228370189667, "reward_std": 0.43745511770248413, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32234376668930054, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.46562498807907104, "step": 157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 402.4375, "completions/mean_terminated_length": 402.4375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.0316, "grad_norm": 0.06490659713745117, "learning_rate": 6.28e-07, "loss": -0.0, "num_tokens": 1852735.0, "reward": 0.4244416356086731, "reward_std": 0.4946388006210327, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16138124465942383, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 432.4375, "completions/mean_terminated_length": 432.4375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0318, "grad_norm": 0.0781712457537651, "learning_rate": 6.319999999999999e-07, "loss": -0.0, "num_tokens": 1863830.0, "reward": 0.5131226778030396, "reward_std": 0.5758332014083862, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27624374628067017, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7918750047683716, "step": 159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 289.0, "completions/max_terminated_length": 289.0, "completions/mean_length": 220.9375, "completions/mean_terminated_length": 220.9375, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 0.032, "grad_norm": 0.048519935458898544, "learning_rate": 6.36e-07, "loss": 0.0, "num_tokens": 1872581.0, "reward": 0.42041048407554626, "reward_std": 0.20900826156139374, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.028318747878074646, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.984375, "step": 160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 830.4375, "completions/mean_terminated_length": 679.888916015625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.0322, "grad_norm": 0.05403288081288338, "learning_rate": 6.4e-07, "loss": 0.0, "num_tokens": 1889756.0, "reward": 0.08898463845252991, "reward_std": 0.2503412663936615, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05296874791383743, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.20625001192092896, "step": 161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 367.1875, "completions/mean_terminated_length": 367.1875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0324, "grad_norm": 0.0862349346280098, "learning_rate": 6.44e-07, "loss": -0.0, "num_tokens": 1899967.0, "reward": 0.40717628598213196, "reward_std": 0.6634520888328552, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2518500089645386, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 555.1875, "completions/mean_terminated_length": 555.1875, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.0326, "grad_norm": 0.059673771262168884, "learning_rate": 6.48e-07, "loss": 0.0, "num_tokens": 1912914.0, "reward": 0.40685755014419556, "reward_std": 0.6638668775558472, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2512125074863434, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8668749928474426, "step": 163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 451.875, "completions/mean_terminated_length": 451.875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0328, "grad_norm": 0.057833489030599594, "learning_rate": 6.52e-07, "loss": 0.0, "num_tokens": 1924224.0, "reward": 0.38958221673965454, "reward_std": 0.5041570663452148, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15416251122951508, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 670.375, "completions/mean_terminated_length": 646.800048828125, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.033, "grad_norm": 0.016138434410095215, "learning_rate": 6.56e-07, "loss": 0.0, "num_tokens": 1939238.0, "reward": 0.09437316656112671, "reward_std": 0.20290447771549225, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0012437499826774001, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9143750667572021, "step": 165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 224.0625, "completions/mean_terminated_length": 224.0625, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0332, "grad_norm": 0.14107482135295868, "learning_rate": 6.6e-07, "loss": 0.0, "num_tokens": 1946823.0, "reward": 0.6916730999946594, "reward_std": 0.6787909269332886, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38334375619888306, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 791.75, "completions/mean_terminated_length": 776.2667236328125, "completions/min_length": 619.0, "completions/min_terminated_length": 619.0, "epoch": 0.0334, "grad_norm": 0.020344126969575882, "learning_rate": 6.64e-07, "loss": 0.0, "num_tokens": 1964547.0, "reward": 0.10289198160171509, "reward_std": 0.22120799124240875, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 414.9375, "completions/mean_terminated_length": 414.9375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.0336, "grad_norm": 0.039657995104789734, "learning_rate": 6.68e-07, "loss": 0.0, "num_tokens": 1975650.0, "reward": 0.23398560285568237, "reward_std": 0.2744271457195282, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 554.0625, "completions/mean_terminated_length": 554.0625, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.0338, "grad_norm": 0.04394695907831192, "learning_rate": 6.72e-07, "loss": 0.0, "num_tokens": 1988195.0, "reward": 0.27687352895736694, "reward_std": 0.4381718933582306, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11624374240636826, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 332.0, "completions/max_terminated_length": 332.0, "completions/mean_length": 246.0, "completions/mean_terminated_length": 246.0, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.034, "grad_norm": 0.1376809924840927, "learning_rate": 6.76e-07, "loss": 0.0, "num_tokens": 1996867.0, "reward": 0.7497512698173523, "reward_std": 0.7743386626243591, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.49950000643730164, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9768750667572021, "step": 170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 626.0, "completions/mean_terminated_length": 599.4666748046875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.0342, "grad_norm": 0.05331079289317131, "learning_rate": 6.800000000000001e-07, "loss": 0.0, "num_tokens": 2010611.0, "reward": 0.21484458446502686, "reward_std": 0.3342306911945343, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1171874925494194, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8237500190734863, "step": 171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 908.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 583.25, "completions/mean_terminated_length": 583.25, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.0344, "grad_norm": 0.04241633415222168, "learning_rate": 6.84e-07, "loss": 0.0, "num_tokens": 2024135.0, "reward": 0.1561727225780487, "reward_std": 0.4363498091697693, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12484374642372131, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.35625001788139343, "step": 172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 363.3125, "completions/mean_terminated_length": 363.3125, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.0346, "grad_norm": 0.036935288459062576, "learning_rate": 6.879999999999999e-07, "loss": 0.0, "num_tokens": 2033756.0, "reward": 0.2983512282371521, "reward_std": 0.2722223103046417, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.034199997782707214, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 564.8125, "completions/mean_terminated_length": 534.2000122070312, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.0348, "grad_norm": 0.05714018642902374, "learning_rate": 6.919999999999999e-07, "loss": -0.0, "num_tokens": 2046857.0, "reward": 0.39007943868637085, "reward_std": 0.5048321485519409, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15515625476837158, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.909375011920929, "step": 174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 509.0625, "completions/mean_terminated_length": 474.7333679199219, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.035, "grad_norm": 0.035533372312784195, "learning_rate": 6.959999999999999e-07, "loss": 0.0, "num_tokens": 2059290.0, "reward": 0.19937589764595032, "reward_std": 0.32233768701553345, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08624999225139618, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.5862500071525574, "step": 175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 294.0625, "completions/mean_terminated_length": 294.0625, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.0352, "grad_norm": 0.08055703341960907, "learning_rate": 7e-07, "loss": 0.0, "num_tokens": 2067915.0, "reward": 0.40953201055526733, "reward_std": 0.44093450903892517, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3190625011920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 221.8125, "completions/mean_terminated_length": 221.8125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.0354, "grad_norm": 0.15087221562862396, "learning_rate": 7.04e-07, "loss": -0.0, "num_tokens": 2074952.0, "reward": 0.6988729238510132, "reward_std": 0.5981594920158386, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33524376153945923, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 270.3125, "completions/mean_terminated_length": 270.3125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.0356, "grad_norm": 0.1314762830734253, "learning_rate": 7.079999999999999e-07, "loss": -0.0, "num_tokens": 2082525.0, "reward": 0.9371042847633362, "reward_std": 0.7496826648712158, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6242062449455261, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 367.3125, "completions/mean_terminated_length": 367.3125, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.0358, "grad_norm": 0.04170224815607071, "learning_rate": 7.119999999999999e-07, "loss": 0.0, "num_tokens": 2092322.0, "reward": 0.100466787815094, "reward_std": 0.21614934504032135, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.013431249186396599, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 394.9375, "completions/mean_terminated_length": 353.0000305175781, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.036, "grad_norm": 0.061775702983140945, "learning_rate": 7.159999999999999e-07, "loss": 0.0, "num_tokens": 2103857.0, "reward": 0.35834813117980957, "reward_std": 0.4030178487300873, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09169375896453857, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9187500476837158, "step": 180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 443.9375, "completions/mean_terminated_length": 443.9375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0362, "grad_norm": 0.05712849646806717, "learning_rate": 7.2e-07, "loss": 0.0, "num_tokens": 2115136.0, "reward": 0.25359198451042175, "reward_std": 0.5182338953018188, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13218125700950623, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9856250286102295, "step": 181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 482.0, "completions/mean_terminated_length": 482.0, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.0364, "grad_norm": 0.038549795746803284, "learning_rate": 7.24e-07, "loss": 0.0, "num_tokens": 2126928.0, "reward": 0.3740636706352234, "reward_std": 0.2613137662410736, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.060624998062849045, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 370.6875, "completions/mean_terminated_length": 370.6875, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.0366, "grad_norm": 0.056822776794433594, "learning_rate": 7.28e-07, "loss": 0.0, "num_tokens": 2136907.0, "reward": 0.6581259369850159, "reward_std": 0.36130666732788086, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2537499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.921875, "step": 183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 313.3125, "completions/mean_terminated_length": 313.3125, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.0368, "grad_norm": 0.06907065212726593, "learning_rate": 7.319999999999999e-07, "loss": -0.0, "num_tokens": 2146416.0, "reward": 0.3497638404369354, "reward_std": 0.5093141198158264, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13702499866485596, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 536.875, "completions/mean_terminated_length": 536.875, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.037, "grad_norm": 0.09015057981014252, "learning_rate": 7.359999999999999e-07, "loss": 0.0, "num_tokens": 2160430.0, "reward": 0.1876571923494339, "reward_std": 0.43939080834388733, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12531250715255737, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6281249523162842, "step": 185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 563.625, "completions/mean_terminated_length": 532.933349609375, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.0372, "grad_norm": 0.03518691658973694, "learning_rate": 7.4e-07, "loss": 0.0, "num_tokens": 2174792.0, "reward": 0.18662890791893005, "reward_std": 0.2931087613105774, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.060756243765354156, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.6575000286102295, "step": 186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 837.5, "completions/mean_terminated_length": 725.6000366210938, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.0374, "grad_norm": 0.03078632988035679, "learning_rate": 7.44e-07, "loss": 0.0, "num_tokens": 2192176.0, "reward": 0.2635161280632019, "reward_std": 0.3674038052558899, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15203124284744263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.48124998807907104, "step": 187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 576.375, "completions/mean_terminated_length": 576.375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.0376, "grad_norm": 0.05314403027296066, "learning_rate": 7.48e-07, "loss": 0.0, "num_tokens": 2205638.0, "reward": 0.3555479049682617, "reward_std": 0.4410718083381653, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1485937535762787, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 360.0, "completions/mean_terminated_length": 360.0, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0378, "grad_norm": 0.07469441741704941, "learning_rate": 7.52e-07, "loss": -0.0, "num_tokens": 2215494.0, "reward": 0.6339856386184692, "reward_std": 0.6436825394630432, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3304687440395355, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.90625, "step": 189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 642.1875, "completions/mean_terminated_length": 616.7333374023438, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.038, "grad_norm": 0.04852589592337608, "learning_rate": 7.559999999999999e-07, "loss": 0.0, "num_tokens": 2229865.0, "reward": 0.22172018885612488, "reward_std": 0.44801998138427734, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13093750178813934, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7987500429153442, "step": 190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 286.625, "completions/mean_terminated_length": 286.625, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.0382, "grad_norm": 0.12670019268989563, "learning_rate": 7.599999999999999e-07, "loss": -0.0, "num_tokens": 2238723.0, "reward": 0.47742295265197754, "reward_std": 0.5909879207611084, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26734375953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.903124988079071, "step": 191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 596.4375, "completions/mean_terminated_length": 535.357177734375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.0384, "grad_norm": 0.03771430253982544, "learning_rate": 7.64e-07, "loss": 0.0, "num_tokens": 2251898.0, "reward": 0.26334771513938904, "reward_std": 0.4172389805316925, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08919374644756317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7743750214576721, "step": 192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 276.3125, "completions/mean_terminated_length": 276.3125, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.0386, "grad_norm": 0.10568352043628693, "learning_rate": 7.68e-07, "loss": 0.0, "num_tokens": 2260223.0, "reward": 0.474116712808609, "reward_std": 0.5619403123855591, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19823125004768372, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 643.5, "completions/mean_terminated_length": 643.5, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.0388, "grad_norm": 0.027317147701978683, "learning_rate": 7.72e-07, "loss": -0.0, "num_tokens": 2274567.0, "reward": 0.20031026005744934, "reward_std": 0.2674041986465454, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.025618748739361763, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7681249380111694, "step": 194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 605.625, "completions/mean_terminated_length": 605.625, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.039, "grad_norm": 0.04695259779691696, "learning_rate": 7.76e-07, "loss": -0.0, "num_tokens": 2287953.0, "reward": 0.23093849420547485, "reward_std": 0.32672035694122314, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08687499910593033, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 353.125, "completions/mean_terminated_length": 353.125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0392, "grad_norm": 0.11250099539756775, "learning_rate": 7.799999999999999e-07, "loss": 0.0, "num_tokens": 2297699.0, "reward": 1.0052103996276855, "reward_std": 0.6778742671012878, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6354187726974487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 798.3125, "completions/mean_terminated_length": 746.2307739257812, "completions/min_length": 561.0, "completions/min_terminated_length": 561.0, "epoch": 0.0394, "grad_norm": 0.02201041951775551, "learning_rate": 7.84e-07, "loss": 0.0, "num_tokens": 2314888.0, "reward": 0.13171659409999847, "reward_std": 0.23589815199375153, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.013431249186396599, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7825000286102295, "step": 197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 399.875, "completions/mean_terminated_length": 399.875, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.0396, "grad_norm": 0.0944603681564331, "learning_rate": 7.88e-07, "loss": 0.0, "num_tokens": 2324950.0, "reward": 0.8473044037818909, "reward_std": 0.6985157132148743, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5071062445640564, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 755.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 451.5625, "completions/mean_terminated_length": 451.5625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0398, "grad_norm": 0.05856088548898697, "learning_rate": 7.92e-07, "loss": -0.0, "num_tokens": 2336223.0, "reward": 0.3316292464733124, "reward_std": 0.36733558773994446, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16325624287128448, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8050000071525574, "step": 199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 499.3125, "completions/mean_terminated_length": 499.3125, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.04, "grad_norm": 0.06026194244623184, "learning_rate": 7.96e-07, "loss": -0.0, "num_tokens": 2347796.0, "reward": 0.20234502851963043, "reward_std": 0.4163602590560913, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09218750149011612, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 332.0, "completions/max_terminated_length": 332.0, "completions/mean_length": 281.5, "completions/mean_terminated_length": 281.5, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.0402, "grad_norm": 0.10090423375368118, "learning_rate": 8e-07, "loss": -0.0, "num_tokens": 2356220.0, "reward": 0.6970885992050171, "reward_std": 0.6772559881210327, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3941749930381775, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.96875, "step": 201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 378.8125, "completions/mean_terminated_length": 378.8125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0404, "grad_norm": 0.04657011851668358, "learning_rate": 8.04e-07, "loss": -0.0, "num_tokens": 2366217.0, "reward": 0.2852294445037842, "reward_std": 0.31723466515541077, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07045625150203705, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8825000524520874, "step": 202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 395.4375, "completions/mean_terminated_length": 395.4375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.0406, "grad_norm": 0.09325200319290161, "learning_rate": 8.08e-07, "loss": 0.0, "num_tokens": 2376528.0, "reward": 0.6974104046821594, "reward_std": 0.6767078042030334, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3948187530040741, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 958.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 615.0625, "completions/mean_terminated_length": 615.0625, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.0408, "grad_norm": 0.04056501388549805, "learning_rate": 8.12e-07, "loss": 0.0, "num_tokens": 2391217.0, "reward": 0.32156357169151306, "reward_std": 0.35403525829315186, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1431249976158142, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.765625, "step": 204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 304.625, "completions/mean_terminated_length": 304.625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.041, "grad_norm": 0.05488446354866028, "learning_rate": 8.159999999999999e-07, "loss": 0.0, "num_tokens": 2400347.0, "reward": 0.4576480984687805, "reward_std": 0.3638550043106079, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10279374569654465, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9718749523162842, "step": 205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 446.4375, "completions/mean_terminated_length": 446.4375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.0412, "grad_norm": 0.039945270866155624, "learning_rate": 8.199999999999999e-07, "loss": -0.0, "num_tokens": 2411442.0, "reward": 0.2057822048664093, "reward_std": 0.27437499165534973, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03656249865889549, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 555.375, "completions/mean_terminated_length": 488.4285888671875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0414, "grad_norm": 0.041412003338336945, "learning_rate": 8.24e-07, "loss": -0.0, "num_tokens": 2424536.0, "reward": 0.2225790172815323, "reward_std": 0.3040996789932251, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07015624642372131, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7056249976158142, "step": 207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 350.75, "completions/mean_terminated_length": 350.75, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.0416, "grad_norm": 0.03908694162964821, "learning_rate": 8.28e-07, "loss": -0.0, "num_tokens": 2434068.0, "reward": 0.2590482234954834, "reward_std": 0.26840072870254517, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01809374988079071, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 417.1875, "completions/mean_terminated_length": 417.1875, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.0418, "grad_norm": 0.06417989730834961, "learning_rate": 8.319999999999999e-07, "loss": -0.0, "num_tokens": 2445511.0, "reward": 0.6460165977478027, "reward_std": 0.5577043890953064, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2920312285423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 571.5, "completions/mean_terminated_length": 541.3333740234375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.042, "grad_norm": 0.02071230858564377, "learning_rate": 8.359999999999999e-07, "loss": -0.0, "num_tokens": 2458479.0, "reward": 0.06859465688467026, "reward_std": 0.18743382394313812, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.8456250429153442, "step": 210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 553.3125, "completions/mean_terminated_length": 553.3125, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.0422, "grad_norm": 0.04340653121471405, "learning_rate": 8.399999999999999e-07, "loss": 0.0, "num_tokens": 2472052.0, "reward": 0.15718816220760345, "reward_std": 0.3381417691707611, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12687499821186066, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.5375000238418579, "step": 211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 351.625, "completions/mean_terminated_length": 351.625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.0424, "grad_norm": 0.07067013531923294, "learning_rate": 8.439999999999999e-07, "loss": 0.0, "num_tokens": 2481710.0, "reward": 0.35882920026779175, "reward_std": 0.5123808979988098, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15515625476837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9612500667572021, "step": 212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 479.6875, "completions/mean_terminated_length": 479.6875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.0426, "grad_norm": 0.06013473495841026, "learning_rate": 8.48e-07, "loss": 0.0, "num_tokens": 2493177.0, "reward": 0.19640758633613586, "reward_std": 0.4110308289527893, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08031249791383743, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 344.75, "completions/mean_terminated_length": 344.75, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.0428, "grad_norm": 0.04810011386871338, "learning_rate": 8.52e-07, "loss": -0.0, "num_tokens": 2502405.0, "reward": 0.638985276222229, "reward_std": 0.3633298873901367, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21546873450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.940625011920929, "step": 214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 448.6875, "completions/mean_terminated_length": 448.6875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.043, "grad_norm": 0.08098689466714859, "learning_rate": 8.559999999999999e-07, "loss": -0.0, "num_tokens": 2513504.0, "reward": 0.8206133842468262, "reward_std": 0.6303836703300476, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4537249803543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8550000190734863, "step": 215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 254.0625, "completions/mean_terminated_length": 254.0625, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.0432, "grad_norm": 0.14319497346878052, "learning_rate": 8.599999999999999e-07, "loss": -0.0, "num_tokens": 2521553.0, "reward": 0.8181759119033813, "reward_std": 0.724412202835083, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5113500356674194, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 756.625, "completions/mean_terminated_length": 718.4285888671875, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.0434, "grad_norm": 0.03249720111489296, "learning_rate": 8.639999999999999e-07, "loss": -0.0, "num_tokens": 2537883.0, "reward": 0.10187577456235886, "reward_std": 0.2792362570762634, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07874999940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.590624988079071, "step": 217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 362.5625, "completions/mean_terminated_length": 362.5625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.0436, "grad_norm": 0.11196468025445938, "learning_rate": 8.68e-07, "loss": 0.0, "num_tokens": 2547300.0, "reward": 0.7413948178291321, "reward_std": 0.7261565327644348, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48278751969337463, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 487.625, "completions/mean_terminated_length": 451.86669921875, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.0438, "grad_norm": 0.06603629142045975, "learning_rate": 8.72e-07, "loss": -0.0, "num_tokens": 2559726.0, "reward": 0.29640740156173706, "reward_std": 0.4160269796848297, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09281250089406967, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.84375, "step": 219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 638.5, "completions/mean_terminated_length": 612.800048828125, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.044, "grad_norm": 0.040623802691698074, "learning_rate": 8.76e-07, "loss": 0.0, "num_tokens": 2573846.0, "reward": 0.4930354058742523, "reward_std": 0.46657171845436096, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1735687404870987, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 252.25, "completions/mean_terminated_length": 252.25, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.0442, "grad_norm": 0.1419198364019394, "learning_rate": 8.799999999999999e-07, "loss": -0.0, "num_tokens": 2581610.0, "reward": 0.7318384647369385, "reward_std": 0.652642011642456, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40117499232292175, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 700.375, "completions/mean_terminated_length": 678.800048828125, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.0444, "grad_norm": 0.017280060797929764, "learning_rate": 8.839999999999999e-07, "loss": -0.0, "num_tokens": 2597232.0, "reward": 0.07148537039756775, "reward_std": 0.19551490247249603, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01796874962747097, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.8162500262260437, "step": 222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 345.125, "completions/mean_terminated_length": 345.125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0446, "grad_norm": 0.12144012749195099, "learning_rate": 8.88e-07, "loss": 0.0, "num_tokens": 2606930.0, "reward": 0.4685792624950409, "reward_std": 0.7178071141242981, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3121562600135803, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 673.5625, "completions/mean_terminated_length": 592.6923217773438, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.0448, "grad_norm": 0.0364852175116539, "learning_rate": 8.92e-07, "loss": -0.0, "num_tokens": 2621707.0, "reward": 0.18093788623809814, "reward_std": 0.38927367329597473, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1743749976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.33937501907348633, "step": 224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 468.125, "completions/mean_terminated_length": 468.125, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.045, "grad_norm": 0.0271297637373209, "learning_rate": 8.96e-07, "loss": 0.0, "num_tokens": 2632909.0, "reward": 0.131716787815094, "reward_std": 0.23589804768562317, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.013431249186396599, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 637.375, "completions/mean_terminated_length": 637.375, "completions/min_length": 515.0, "completions/min_terminated_length": 515.0, "epoch": 0.0452, "grad_norm": 0.033811796456575394, "learning_rate": 9e-07, "loss": 0.0, "num_tokens": 2647283.0, "reward": 0.09374824166297913, "reward_std": 0.37498739361763, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06249374896287918, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 895.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 624.0, "completions/mean_terminated_length": 624.0, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.0454, "grad_norm": 0.04741629213094711, "learning_rate": 9.039999999999999e-07, "loss": 0.0, "num_tokens": 2662259.0, "reward": 0.09648525714874268, "reward_std": 0.27594712376594543, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06796874850988388, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.5, "step": 227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 515.375, "completions/mean_terminated_length": 515.375, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.0456, "grad_norm": 0.07952979952096939, "learning_rate": 9.08e-07, "loss": 0.0, "num_tokens": 2674537.0, "reward": 0.6585164070129395, "reward_std": 0.5871793031692505, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5045312643051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.703125, "step": 228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 352.25, "completions/mean_terminated_length": 352.25, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.0458, "grad_norm": 0.08358965814113617, "learning_rate": 9.12e-07, "loss": -0.0, "num_tokens": 2685613.0, "reward": 0.5863288640975952, "reward_std": 0.4692271947860718, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.42265626788139343, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6375000476837158, "step": 229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 532.3125, "completions/mean_terminated_length": 532.3125, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.046, "grad_norm": 0.03617256507277489, "learning_rate": 9.16e-07, "loss": 0.0, "num_tokens": 2698002.0, "reward": 0.3411603271961212, "reward_std": 0.27301958203315735, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05731874704360962, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 567.875, "completions/mean_terminated_length": 567.875, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.0462, "grad_norm": 0.04010525345802307, "learning_rate": 9.2e-07, "loss": -0.0, "num_tokens": 2710944.0, "reward": 0.2974105179309845, "reward_std": 0.41598159074783325, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09481874853372574, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 391.5, "completions/mean_terminated_length": 391.5, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0464, "grad_norm": 0.04272209107875824, "learning_rate": 9.24e-07, "loss": 0.0, "num_tokens": 2721416.0, "reward": 0.19850128889083862, "reward_std": 0.2649759352207184, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02199999988079071, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 244.875, "completions/mean_terminated_length": 244.875, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.0466, "grad_norm": 0.1290624886751175, "learning_rate": 9.28e-07, "loss": 0.0, "num_tokens": 2729206.0, "reward": 1.0090575218200684, "reward_std": 0.5970871448516846, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5806125402450562, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 603.0, "completions/mean_terminated_length": 603.0, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.0468, "grad_norm": 0.045601651072502136, "learning_rate": 9.32e-07, "loss": 0.0, "num_tokens": 2743078.0, "reward": 0.15717624127864838, "reward_std": 0.35260462760925293, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12685000896453857, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.5518749952316284, "step": 234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 382.9375, "completions/mean_terminated_length": 382.9375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.047, "grad_norm": 0.09367775917053223, "learning_rate": 9.36e-07, "loss": -0.0, "num_tokens": 2753669.0, "reward": 0.6351448893547058, "reward_std": 0.5878840684890747, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33278748393058777, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9206249713897705, "step": 235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 788.9375, "completions/mean_terminated_length": 734.6923217773438, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.0472, "grad_norm": 0.020629653707146645, "learning_rate": 9.399999999999999e-07, "loss": -0.0, "num_tokens": 2770452.0, "reward": 0.0843757912516594, "reward_std": 0.2323996126651764, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04374999925494194, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6687500476837158, "step": 236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 237.0, "completions/max_terminated_length": 237.0, "completions/mean_length": 183.625, "completions/mean_terminated_length": 183.625, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.0474, "grad_norm": 0.12108368426561356, "learning_rate": 9.439999999999999e-07, "loss": 0.0, "num_tokens": 2777166.0, "reward": 1.311863899230957, "reward_std": 0.5121001601219177, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8737249970436096, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9781249761581421, "step": 237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 730.6875, "completions/mean_terminated_length": 711.1333618164062, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.0476, "grad_norm": 0.03740872070193291, "learning_rate": 9.479999999999999e-07, "loss": 0.0, "num_tokens": 2793609.0, "reward": 0.19294483959674835, "reward_std": 0.40760478377342224, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0733874961733818, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8487500548362732, "step": 238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 747.9375, "completions/mean_terminated_length": 684.2307739257812, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.0478, "grad_norm": 0.023466261103749275, "learning_rate": 9.52e-07, "loss": 0.0, "num_tokens": 2809624.0, "reward": 0.09984464198350906, "reward_std": 0.21490369737148285, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.5718750357627869, "step": 239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 387.8125, "completions/mean_terminated_length": 387.8125, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.048, "grad_norm": 0.0626145526766777, "learning_rate": 9.559999999999998e-07, "loss": 0.0, "num_tokens": 2819781.0, "reward": 0.3981103301048279, "reward_std": 0.3956868350505829, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10871875286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 821.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 579.8125, "completions/mean_terminated_length": 579.8125, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.0482, "grad_norm": 0.05034630745649338, "learning_rate": 9.6e-07, "loss": 0.0, "num_tokens": 2833874.0, "reward": 0.21937349438667297, "reward_std": 0.5157561302185059, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1262437403202057, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9936875104904175, "step": 241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 287.625, "completions/mean_terminated_length": 287.625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.0484, "grad_norm": 0.054451264441013336, "learning_rate": 9.64e-07, "loss": -0.0, "num_tokens": 2843308.0, "reward": 0.38210445642471313, "reward_std": 0.27248647809028625, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07670624554157257, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 372.0, "completions/max_terminated_length": 372.0, "completions/mean_length": 246.875, "completions/mean_terminated_length": 246.875, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.0486, "grad_norm": 0.0603388175368309, "learning_rate": 9.679999999999999e-07, "loss": 0.0, "num_tokens": 2851962.0, "reward": 0.2983948588371277, "reward_std": 0.27209776639938354, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03428749740123749, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8506250381469727, "step": 243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 513.6875, "completions/mean_terminated_length": 479.66668701171875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.0488, "grad_norm": 0.03866273909807205, "learning_rate": 9.72e-07, "loss": -0.0, "num_tokens": 2864261.0, "reward": 0.15804781019687653, "reward_std": 0.35231274366378784, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12859374284744263, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.625, "step": 244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 680.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 488.625, "completions/mean_terminated_length": 488.625, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.049, "grad_norm": 0.0689377710223198, "learning_rate": 9.759999999999998e-07, "loss": 0.0, "num_tokens": 2876351.0, "reward": 0.42137908935546875, "reward_std": 0.4937266409397125, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1552562415599823, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 335.0, "completions/mean_terminated_length": 335.0, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.0492, "grad_norm": 0.06953722983598709, "learning_rate": 9.8e-07, "loss": 0.0, "num_tokens": 2886175.0, "reward": 0.490588515996933, "reward_std": 0.4661201536655426, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1686749905347824, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 291.6875, "completions/mean_terminated_length": 291.6875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.0494, "grad_norm": 0.09857507050037384, "learning_rate": 9.84e-07, "loss": 0.0, "num_tokens": 2894634.0, "reward": 0.6652479767799377, "reward_std": 0.6224967241287231, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3304937481880188, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.971875011920929, "step": 247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 653.0625, "completions/mean_terminated_length": 628.3333740234375, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.0496, "grad_norm": 0.04875617101788521, "learning_rate": 9.88e-07, "loss": -0.0, "num_tokens": 2908731.0, "reward": 0.3392167389392853, "reward_std": 0.5302200317382812, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17843125760555267, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8274999856948853, "step": 248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 574.4375, "completions/mean_terminated_length": 574.4375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.0498, "grad_norm": 0.08131159842014313, "learning_rate": 9.92e-07, "loss": -0.0, "num_tokens": 2922034.0, "reward": 0.5370946526527405, "reward_std": 0.6940097212791443, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3241874873638153, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 299.0, "completions/mean_terminated_length": 299.0, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.05, "grad_norm": 0.1352805495262146, "learning_rate": 9.959999999999999e-07, "loss": -0.0, "num_tokens": 2930626.0, "reward": 0.877319872379303, "reward_std": 0.7394229769706726, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5671374797821045, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 869.0, "completions/max_terminated_length": 869.0, "completions/mean_length": 577.375, "completions/mean_terminated_length": 577.375, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.0502, "grad_norm": 0.05928274244070053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2943928.0, "reward": 0.5016261339187622, "reward_std": 0.6571268439292908, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31575000286102295, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9256250262260437, "step": 251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 480.375, "completions/mean_terminated_length": 480.375, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.0504, "grad_norm": 0.08286119252443314, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2955630.0, "reward": 0.6260949969291687, "reward_std": 0.663368284702301, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37718749046325684, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.878125011920929, "step": 252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 705.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 368.125, "completions/mean_terminated_length": 368.125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.0506, "grad_norm": 0.0810019001364708, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2965424.0, "reward": 0.5696882009506226, "reward_std": 0.5016130805015564, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3268750011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7549999952316284, "step": 253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 583.1875, "completions/mean_terminated_length": 553.800048828125, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.0508, "grad_norm": 0.03166787326335907, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2981235.0, "reward": 0.14007914066314697, "reward_std": 0.25079262256622314, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.030156249180436134, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 565.1875, "completions/mean_terminated_length": 565.1875, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.051, "grad_norm": 0.05701864883303642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2994646.0, "reward": 0.32656025886535645, "reward_std": 0.398924857378006, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21561875939369202, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.6368750333786011, "step": 255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 496.9375, "completions/mean_terminated_length": 496.9375, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.0512, "grad_norm": 0.09092193096876144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3006885.0, "reward": 0.5965479016304016, "reward_std": 0.7346311211585999, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3805937170982361, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 306.0625, "completions/mean_terminated_length": 306.0625, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0514, "grad_norm": 0.11935282498598099, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3016006.0, "reward": 0.7588229775428772, "reward_std": 0.6787295937538147, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5176437497138977, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8243750333786011, "step": 257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 411.5625, "completions/mean_terminated_length": 411.5625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.0516, "grad_norm": 0.054954513907432556, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3026607.0, "reward": 0.26701968908309937, "reward_std": 0.45922335982322693, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1590375006198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8256250023841858, "step": 258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 331.75, "completions/mean_terminated_length": 331.75, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.0518, "grad_norm": 0.09276754409074783, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3036091.0, "reward": 0.7678916454315186, "reward_std": 0.6246877312660217, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41078123450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 623.3125, "completions/mean_terminated_length": 566.0714721679688, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.052, "grad_norm": 0.11961908638477325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3049920.0, "reward": 0.6557040214538574, "reward_std": 0.7678809762001038, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4364062547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8149999380111694, "step": 260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 465.75, "completions/mean_terminated_length": 428.5333557128906, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.0522, "grad_norm": 0.03776947781443596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3061660.0, "reward": 0.1347634643316269, "reward_std": 0.24122655391693115, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01952499896287918, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 352.6875, "completions/mean_terminated_length": 352.6875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0524, "grad_norm": 0.0890905037522316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3071735.0, "reward": 0.5494418144226074, "reward_std": 0.5305500626564026, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22388124465942383, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 260.4375, "completions/mean_terminated_length": 260.4375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.0526, "grad_norm": 0.09029641002416611, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3079966.0, "reward": 0.6812385320663452, "reward_std": 0.5318145751953125, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2999749779701233, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9393750429153442, "step": 263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 491.5625, "completions/mean_terminated_length": 491.5625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0528, "grad_norm": 0.07237495481967926, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3091943.0, "reward": 0.5379419326782227, "reward_std": 0.618116557598114, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2633812427520752, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 402.125, "completions/mean_terminated_length": 402.125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.053, "grad_norm": 0.08158623427152634, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3102153.0, "reward": 0.9773197174072266, "reward_std": 0.6373648047447205, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5796375274658203, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8974999785423279, "step": 265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 377.6875, "completions/mean_terminated_length": 377.6875, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0532, "grad_norm": 0.103713758289814, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3111876.0, "reward": 0.6279606819152832, "reward_std": 0.7182199954986572, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3809187412261963, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9075000286102295, "step": 266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 497.5625, "completions/mean_terminated_length": 497.5625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.0534, "grad_norm": 0.08784513175487518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3124045.0, "reward": 0.36140719056129456, "reward_std": 0.5140721797943115, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1603125035762787, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8343750238418579, "step": 267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 660.625, "completions/mean_terminated_length": 608.7142944335938, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.0536, "grad_norm": 0.03890775516629219, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3138439.0, "reward": 0.5287507176399231, "reward_std": 0.39698562026023865, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24500000476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 861.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 511.6875, "completions/mean_terminated_length": 511.6875, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.0538, "grad_norm": 0.04824133589863777, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3150370.0, "reward": 0.2154664546251297, "reward_std": 0.35149630904197693, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11843124777078629, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.5718749761581421, "step": 269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 473.0625, "completions/mean_terminated_length": 473.0625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.054, "grad_norm": 0.10594906657934189, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3161331.0, "reward": 0.6506197452545166, "reward_std": 0.7182503342628479, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4262375235557556, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 365.0, "completions/mean_terminated_length": 365.0, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.0542, "grad_norm": 0.12241111695766449, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3171139.0, "reward": 0.8522665500640869, "reward_std": 0.6953026056289673, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5170312523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 640.6875, "completions/mean_terminated_length": 585.9285888671875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.0544, "grad_norm": 0.039895884692668915, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3185902.0, "reward": 0.0974227637052536, "reward_std": 0.2711142897605896, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06984374672174454, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.643750011920929, "step": 272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 323.875, "completions/mean_terminated_length": 323.875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0546, "grad_norm": 0.13070128858089447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3195340.0, "reward": 0.5621105432510376, "reward_std": 0.7494791746139526, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.37421876192092896, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 572.5625, "completions/mean_terminated_length": 572.5625, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.0548, "grad_norm": 0.047018591314554214, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3208885.0, "reward": 0.19695085287094116, "reward_std": 0.3196689784526825, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08139999955892563, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.7300000190734863, "step": 274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 807.5625, "completions/mean_terminated_length": 776.6428833007812, "completions/min_length": 541.0, "completions/min_terminated_length": 541.0, "epoch": 0.055, "grad_norm": 0.02187122218310833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3226318.0, "reward": 0.1399228274822235, "reward_std": 0.2511211633682251, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.029843749478459358, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7625000476837158, "step": 275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 294.25, "completions/mean_terminated_length": 294.25, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.0552, "grad_norm": 0.10627247393131256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3234770.0, "reward": 0.8775542974472046, "reward_std": 0.7396323680877686, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5676062107086182, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 591.0625, "completions/mean_terminated_length": 529.2142944335938, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.0554, "grad_norm": 0.09432274848222733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3249043.0, "reward": 0.37975114583969116, "reward_std": 0.5925374627113342, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19699999690055847, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8512499928474426, "step": 277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 709.25, "completions/mean_terminated_length": 688.2667236328125, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.0556, "grad_norm": 0.03390571475028992, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3264295.0, "reward": 0.10843874514102936, "reward_std": 0.2991413474082947, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09187500178813934, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 999.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 621.0625, "completions/mean_terminated_length": 621.0625, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.0558, "grad_norm": 0.05371436849236488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3278632.0, "reward": 0.31538859009742737, "reward_std": 0.6028802990913391, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19327500462532043, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.90625, "step": 279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 625.9375, "completions/mean_terminated_length": 625.9375, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.056, "grad_norm": 0.05514480173587799, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3292983.0, "reward": 0.2764856815338135, "reward_std": 0.538123607635498, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17796874046325684, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 768.6875, "completions/mean_terminated_length": 709.769287109375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.0562, "grad_norm": 0.044265639036893845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3309474.0, "reward": 0.19539141654968262, "reward_std": 0.3165375590324402, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07828124612569809, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.4656250476837158, "step": 281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 421.75, "completions/mean_terminated_length": 421.75, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0564, "grad_norm": 0.06593500077724457, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3320206.0, "reward": 0.5497353076934814, "reward_std": 0.46284019947052, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22446876764297485, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 500.0, "completions/mean_terminated_length": 465.0666809082031, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.0566, "grad_norm": 0.07852934300899506, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3332142.0, "reward": 0.4702320992946625, "reward_std": 0.5841633677482605, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25296249985694885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8168749809265137, "step": 283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 696.0, "completions/mean_terminated_length": 674.1333618164062, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.0568, "grad_norm": 0.05826333910226822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3347518.0, "reward": 0.3153793513774872, "reward_std": 0.6028605699539185, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19325625896453857, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7893750071525574, "step": 284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 267.1875, "completions/mean_terminated_length": 267.1875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.057, "grad_norm": 0.04542062431573868, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3355601.0, "reward": 0.442813515663147, "reward_std": 0.22002913057804108, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07312499731779099, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.953125, "step": 285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 335.5625, "completions/mean_terminated_length": 335.5625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0572, "grad_norm": 0.12789709866046906, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3365178.0, "reward": 1.099510669708252, "reward_std": 0.6321114301681519, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6990187764167786, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 519.0, "completions/mean_terminated_length": 485.3333435058594, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.0574, "grad_norm": 0.027280030772089958, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3378426.0, "reward": 0.12562355399131775, "reward_std": 0.22473031282424927, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0012437499826774001, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 553.3125, "completions/mean_terminated_length": 553.3125, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.0576, "grad_norm": 0.03381048142910004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3391791.0, "reward": 0.20578226447105408, "reward_std": 0.27437496185302734, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03656249865889549, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8862500190734863, "step": 288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 335.3125, "completions/mean_terminated_length": 335.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.0578, "grad_norm": 0.1040911078453064, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3400868.0, "reward": 0.5461728572845459, "reward_std": 0.6174020767211914, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2798437476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 356.8125, "completions/mean_terminated_length": 356.8125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.058, "grad_norm": 0.0443212166428566, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3410081.0, "reward": 0.4082043468952179, "reward_std": 0.24492977559566498, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06640625, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 404.5625, "completions/mean_terminated_length": 404.5625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.0582, "grad_norm": 0.06634201854467392, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3420474.0, "reward": 0.42742300033569336, "reward_std": 0.49513331055641174, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16734375059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 645.9375, "completions/mean_terminated_length": 645.9375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.0584, "grad_norm": 0.04183226451277733, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3434761.0, "reward": 0.25725406408309937, "reward_std": 0.45165103673934937, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1395062506198883, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.7737499475479126, "step": 292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 400.9375, "completions/mean_terminated_length": 400.9375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.0586, "grad_norm": 0.04983692988753319, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3445080.0, "reward": 0.12616072595119476, "reward_std": 0.38831111788749695, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06481874734163284, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9137499928474426, "step": 293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 816.625, "completions/mean_terminated_length": 768.769287109375, "completions/min_length": 527.0, "completions/min_terminated_length": 527.0, "epoch": 0.0588, "grad_norm": 0.033462971448898315, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3463042.0, "reward": 0.19328199326992035, "reward_std": 0.36465102434158325, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1365625113248825, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5618749856948853, "step": 294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 839.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 542.4375, "completions/mean_terminated_length": 542.4375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.059, "grad_norm": 0.03829784691333771, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3475689.0, "reward": 0.37664180994033813, "reward_std": 0.2644560635089874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06578125059604645, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.953125, "step": 295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 375.5, "completions/mean_terminated_length": 375.5, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.0592, "grad_norm": 0.0811561867594719, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3485745.0, "reward": 0.4990634024143219, "reward_std": 0.4820137619972229, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.43562501668930054, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.39374998211860657, "step": 296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 427.5, "completions/mean_terminated_length": 427.5, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.0594, "grad_norm": 0.08680995553731918, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3497337.0, "reward": 0.4759261906147003, "reward_std": 0.645025908946991, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2643499970436096, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9643749594688416, "step": 297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 649.375, "completions/mean_terminated_length": 595.857177734375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.0596, "grad_norm": 0.057836707681417465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3511903.0, "reward": 0.2187509834766388, "reward_std": 0.48196664452552795, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1875, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6825000047683716, "step": 298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 739.625, "completions/mean_terminated_length": 720.6666870117188, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.0598, "grad_norm": 0.03151189535856247, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3528649.0, "reward": 0.1990632265806198, "reward_std": 0.31180843710899353, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08562499284744263, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.6681249737739563, "step": 299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 252.1875, "completions/mean_terminated_length": 252.1875, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.06, "grad_norm": 0.12714387476444244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3536972.0, "reward": 1.1587355136871338, "reward_std": 0.6213690638542175, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7549687623977661, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 488.5, "completions/mean_terminated_length": 488.5, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.0602, "grad_norm": 0.06142599135637283, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3548836.0, "reward": 0.45867300033569336, "reward_std": 0.4819618761539459, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16734375059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 548.4375, "completions/mean_terminated_length": 548.4375, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.0604, "grad_norm": 0.04295762628316879, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3561435.0, "reward": 0.3402353525161743, "reward_std": 0.3167971670627594, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11796875298023224, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7862499952316284, "step": 302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 668.375, "completions/mean_terminated_length": 668.375, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.0606, "grad_norm": 0.024089492857456207, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3576641.0, "reward": 0.0371885672211647, "reward_std": 0.1487499624490738, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.011874999850988388, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 469.0, "completions/mean_terminated_length": 469.0, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.0608, "grad_norm": 0.037190888077020645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3588401.0, "reward": 0.13414177298545837, "reward_std": 0.2402055412530899, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 441.125, "completions/mean_terminated_length": 441.125, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.061, "grad_norm": 0.03902782127261162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3599123.0, "reward": 0.5672666430473328, "reward_std": 0.21187499165534973, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19703125953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 358.3125, "completions/mean_terminated_length": 358.3125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0612, "grad_norm": 0.08498743921518326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3608856.0, "reward": 0.9171764254570007, "reward_std": 0.5459103584289551, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45934998989105225, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.984375, "step": 306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 255.0625, "completions/mean_terminated_length": 255.0625, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.0614, "grad_norm": 0.08077023178339005, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3616649.0, "reward": 0.507960855960846, "reward_std": 0.36551669239997864, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14091874659061432, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 262.0, "completions/max_terminated_length": 262.0, "completions/mean_length": 201.8125, "completions/mean_terminated_length": 201.8125, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 0.0616, "grad_norm": 0.11487986892461777, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3624198.0, "reward": 1.2167980670928955, "reward_std": 0.6037032604217529, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.80859375, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.949999988079071, "step": 308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 301.9375, "completions/mean_terminated_length": 301.9375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.0618, "grad_norm": 0.1313762366771698, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3633029.0, "reward": 1.1242854595184326, "reward_std": 0.6703938841819763, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7485687732696533, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 478.0625, "completions/mean_terminated_length": 478.0625, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.062, "grad_norm": 0.07958004623651505, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3644870.0, "reward": 0.5756134986877441, "reward_std": 0.600468099117279, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2762250006198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8675000071525574, "step": 310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 424.25, "completions/mean_terminated_length": 424.25, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0622, "grad_norm": 0.09165378659963608, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3655946.0, "reward": 0.5120260715484619, "reward_std": 0.6325606107711792, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2740499973297119, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 698.0, "completions/max_terminated_length": 698.0, "completions/mean_length": 509.6875, "completions/mean_terminated_length": 509.6875, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.0624, "grad_norm": 0.09163699299097061, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3667685.0, "reward": 0.817969799041748, "reward_std": 0.7242051959037781, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.510937511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 675.1875, "completions/mean_terminated_length": 675.1875, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.0626, "grad_norm": 0.044196583330631256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3682472.0, "reward": 0.1586725413799286, "reward_std": 0.3538687527179718, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12984374165534973, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.4781250059604645, "step": 313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 266.375, "completions/mean_terminated_length": 266.375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.0628, "grad_norm": 0.14794600009918213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3691310.0, "reward": 0.7589044570922852, "reward_std": 0.7036590576171875, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4553062319755554, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 926.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 465.375, "completions/mean_terminated_length": 465.375, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.063, "grad_norm": 0.018205508589744568, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3702644.0, "reward": 0.515535295009613, "reward_std": 0.13819099962711334, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09356874227523804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 371.5, "completions/mean_terminated_length": 371.5, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.0632, "grad_norm": 0.04262479394674301, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3712492.0, "reward": 0.17132914066314697, "reward_std": 0.26302751898765564, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.030156249180436134, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 275.875, "completions/mean_terminated_length": 275.875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0634, "grad_norm": 0.14935748279094696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3720714.0, "reward": 1.0056791305541992, "reward_std": 0.6782376766204834, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6363562941551208, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.96875, "step": 317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 260.25, "completions/mean_terminated_length": 260.25, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.0636, "grad_norm": 0.12957580387592316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3728670.0, "reward": 0.9429698586463928, "reward_std": 0.6755355596542358, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.573437511920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 287.625, "completions/mean_terminated_length": 287.625, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.0638, "grad_norm": 0.05074005946516991, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3736776.0, "reward": 0.5671696662902832, "reward_std": 0.28308388590812683, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13433749973773956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 606.25, "completions/mean_terminated_length": 606.25, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.064, "grad_norm": 0.022477490827441216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3750796.0, "reward": 0.0967981144785881, "reward_std": 0.20836085081100464, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.671875, "step": 320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 471.0, "completions/max_terminated_length": 471.0, "completions/mean_length": 358.5625, "completions/mean_terminated_length": 358.5625, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.0642, "grad_norm": 0.09163764119148254, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3760357.0, "reward": 0.8826569318771362, "reward_std": 0.5445376634597778, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5153125524520874, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5249999761581421, "step": 321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 299.6875, "completions/mean_terminated_length": 299.6875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0644, "grad_norm": 0.1221647784113884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3768768.0, "reward": 0.8254446983337402, "reward_std": 0.648068368434906, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46338751912117004, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 357.25, "completions/mean_terminated_length": 357.25, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0646, "grad_norm": 0.06700371950864792, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3778148.0, "reward": 0.4514702558517456, "reward_std": 0.36238211393356323, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09043750166893005, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9200000166893005, "step": 323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 765.3125, "completions/mean_terminated_length": 679.0833740234375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0648, "grad_norm": 0.07601146399974823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3794601.0, "reward": 0.29265692830085754, "reward_std": 0.5199978947639465, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14781250059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.6368749737739563, "step": 324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 419.75, "completions/mean_terminated_length": 419.75, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.065, "grad_norm": 0.0935240164399147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3805509.0, "reward": 0.45671340823173523, "reward_std": 0.5214563608169556, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1634249985218048, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8425000309944153, "step": 325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 626.1875, "completions/mean_terminated_length": 599.6666870117188, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.0652, "grad_norm": 0.030941950157284737, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3819944.0, "reward": 0.16054198145866394, "reward_std": 0.24615062773227692, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.008581249974668026, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9193750023841858, "step": 326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 543.0625, "completions/mean_terminated_length": 511.0000305175781, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.0654, "grad_norm": 0.07703561335802078, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3832441.0, "reward": 0.44718241691589355, "reward_std": 0.5750358700752258, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20686249434947968, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8531250357627869, "step": 327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 756.875, "completions/mean_terminated_length": 718.7142944335938, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.0656, "grad_norm": 0.04318975284695625, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3849431.0, "reward": 0.18656326830387115, "reward_std": 0.40109699964523315, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18562500178813934, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.25, "step": 328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 594.125, "completions/mean_terminated_length": 594.125, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.0658, "grad_norm": 0.07453557848930359, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3863033.0, "reward": 0.4498445391654968, "reward_std": 0.5315169095993042, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2746874988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7775000333786011, "step": 329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 554.75, "completions/mean_terminated_length": 554.75, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.066, "grad_norm": 0.03961712121963501, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3876357.0, "reward": 0.16507934033870697, "reward_std": 0.31170380115509033, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08015625178813934, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 451.125, "completions/mean_terminated_length": 451.125, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.0662, "grad_norm": 0.040022239089012146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3887527.0, "reward": 0.34920740127563477, "reward_std": 0.28192847967147827, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07341250032186508, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9512499570846558, "step": 331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 634.375, "completions/mean_terminated_length": 608.4000244140625, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.0664, "grad_norm": 0.06123507767915726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3901197.0, "reward": 0.28600096702575684, "reward_std": 0.51737380027771, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13449999690055847, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 542.875, "completions/mean_terminated_length": 510.8000183105469, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.0666, "grad_norm": 0.0665750652551651, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3914027.0, "reward": 0.41491949558258057, "reward_std": 0.44320619106292725, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32983750104904175, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.6775000095367432, "step": 333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 444.4375, "completions/mean_terminated_length": 444.4375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.0668, "grad_norm": 0.08560356497764587, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3925202.0, "reward": 0.6976916790008545, "reward_std": 0.6770631670951843, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39538127183914185, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 392.125, "completions/mean_terminated_length": 392.125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.067, "grad_norm": 0.05528002604842186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3935140.0, "reward": 0.7269509434700012, "reward_std": 0.383118599653244, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26639997959136963, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 459.9375, "completions/mean_terminated_length": 422.3333435058594, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.0672, "grad_norm": 0.04702359065413475, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3947075.0, "reward": 0.35108882188796997, "reward_std": 0.30552181601524353, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07717499881982803, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7981249690055847, "step": 336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 532.375, "completions/mean_terminated_length": 532.375, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.0674, "grad_norm": 0.05920146405696869, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3959705.0, "reward": 0.31739193201065063, "reward_std": 0.5143815875053406, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13478124141693115, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 455.5625, "completions/mean_terminated_length": 455.5625, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.0676, "grad_norm": 0.1070626825094223, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3970962.0, "reward": 0.5623135566711426, "reward_std": 0.749750018119812, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.374625027179718, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 841.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 588.0, "completions/mean_terminated_length": 588.0, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.0678, "grad_norm": 0.07955072820186615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3984738.0, "reward": 0.40898579359054565, "reward_std": 0.6638203859329224, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2554687559604645, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.949999988079071, "step": 339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 465.125, "completions/mean_terminated_length": 465.125, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.068, "grad_norm": 0.0848647952079773, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3996036.0, "reward": 0.3072665333747864, "reward_std": 0.4662124514579773, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1770312488079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7593750357627869, "step": 340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 617.1875, "completions/mean_terminated_length": 590.0667114257812, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.0682, "grad_norm": 0.08455904573202133, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4009831.0, "reward": 0.8132666945457458, "reward_std": 0.7238873839378357, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5015312433242798, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 663.6875, "completions/mean_terminated_length": 663.6875, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.0684, "grad_norm": 0.07552959769964218, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4024498.0, "reward": 0.40905115008354187, "reward_std": 0.6639341711997986, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2556000053882599, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8475000262260437, "step": 342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 327.6875, "completions/mean_terminated_length": 327.6875, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.0686, "grad_norm": 0.11827276647090912, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4034221.0, "reward": 0.47278860211372375, "reward_std": 0.6446200013160706, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2580749988555908, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 344.125, "completions/mean_terminated_length": 344.125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0688, "grad_norm": 0.12455980479717255, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4043679.0, "reward": 0.8523322939872742, "reward_std": 0.6953675746917725, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5171625018119812, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 341.3125, "completions/mean_terminated_length": 341.3125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.069, "grad_norm": 0.13736172020435333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4052900.0, "reward": 0.4060136079788208, "reward_std": 0.6633785367012024, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2495250105857849, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 385.9375, "completions/mean_terminated_length": 385.9375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0692, "grad_norm": 0.09651404619216919, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4063059.0, "reward": 0.5610165596008301, "reward_std": 0.5303348898887634, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24703124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 518.0625, "completions/mean_terminated_length": 401.3077087402344, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.0694, "grad_norm": 0.06404157727956772, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4075412.0, "reward": 0.3319413661956787, "reward_std": 0.4131385087966919, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10138124227523804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7737499475479126, "step": 347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 391.25, "completions/mean_terminated_length": 391.25, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0696, "grad_norm": 0.06245727464556694, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4086120.0, "reward": 0.41967636346817017, "reward_std": 0.42084556818008423, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15184998512268066, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8675000071525574, "step": 348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 498.5625, "completions/mean_terminated_length": 498.5625, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.0698, "grad_norm": 0.08765008300542831, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4098273.0, "reward": 0.44343554973602295, "reward_std": 0.6556655764579773, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26186874508857727, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9649999737739563, "step": 349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 340.8125, "completions/mean_terminated_length": 340.8125, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.07, "grad_norm": 0.12874794006347656, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4107998.0, "reward": 0.5621886253356934, "reward_std": 0.7495835423469543, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.37437498569488525, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 435.875, "completions/mean_terminated_length": 435.875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.0702, "grad_norm": 0.09282677620649338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4120204.0, "reward": 0.6560107469558716, "reward_std": 0.7682395577430725, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4370187520980835, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 534.125, "completions/mean_terminated_length": 501.4667053222656, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0704, "grad_norm": 0.10270912200212479, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4132958.0, "reward": 0.5584321022033691, "reward_std": 0.66446852684021, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3043625056743622, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8456249833106995, "step": 352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 247.0, "completions/max_terminated_length": 247.0, "completions/mean_length": 199.375, "completions/mean_terminated_length": 199.375, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.0706, "grad_norm": 0.1691531240940094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4140068.0, "reward": 1.0624885559082031, "reward_std": 0.6790942549705505, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6874749660491943, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 359.625, "completions/mean_terminated_length": 359.625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.0708, "grad_norm": 0.12608271837234497, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4150126.0, "reward": 0.34671685099601746, "reward_std": 0.5985450148582458, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19343124330043793, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 386.3125, "completions/mean_terminated_length": 386.3125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.071, "grad_norm": 0.07373722642660141, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4160419.0, "reward": 0.5789071917533875, "reward_std": 0.459003746509552, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2828125059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.887499988079071, "step": 355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 303.0, "completions/max_terminated_length": 303.0, "completions/mean_length": 244.6875, "completions/mean_terminated_length": 244.6875, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.0712, "grad_norm": 0.14102409780025482, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4168414.0, "reward": 1.0081697702407837, "reward_std": 0.5978062748908997, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.578837513923645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 547.75, "completions/mean_terminated_length": 516.0, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.0714, "grad_norm": 0.06826342642307281, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4180938.0, "reward": 0.4097512364387512, "reward_std": 0.5840522646903992, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19449999928474426, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.7962499856948853, "step": 357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 332.125, "completions/mean_terminated_length": 332.125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0716, "grad_norm": 0.12041150033473969, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4191196.0, "reward": 0.8805197477340698, "reward_std": 0.7384181618690491, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5735375285148621, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 380.75, "completions/mean_terminated_length": 380.75, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0718, "grad_norm": 0.12349335104227066, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4201208.0, "reward": 0.377963662147522, "reward_std": 0.5922492742538452, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1934249997138977, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 343.75, "completions/mean_terminated_length": 343.75, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.072, "grad_norm": 0.08135705441236496, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4211316.0, "reward": 0.3307041525840759, "reward_std": 0.4125674366950989, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0989062488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 701.0, "completions/mean_terminated_length": 679.4666748046875, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.0722, "grad_norm": 0.053534626960754395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4227348.0, "reward": 0.1793757528066635, "reward_std": 0.3858878016471863, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17125000059604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6212500333786011, "step": 361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 451.5, "completions/mean_terminated_length": 413.3333435058594, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.0724, "grad_norm": 0.09563141316175461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4238684.0, "reward": 0.7721759676933289, "reward_std": 0.5366109013557434, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3568499982357025, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 274.0, "completions/mean_terminated_length": 274.0, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.0726, "grad_norm": 0.07545175403356552, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4246972.0, "reward": 0.5620291233062744, "reward_std": 0.42517292499542236, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18655624985694885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 781.375, "completions/mean_terminated_length": 700.5, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0728, "grad_norm": 0.05049436166882515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4264370.0, "reward": 0.29062575101852417, "reward_std": 0.3986178934574127, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20624999701976776, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5525000095367432, "step": 364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 518.0625, "completions/mean_terminated_length": 484.3333435058594, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.073, "grad_norm": 0.09847228229045868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4276659.0, "reward": 0.9992980360984802, "reward_std": 0.6825810074806213, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6235937476158142, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.90625, "step": 365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 274.6875, "completions/mean_terminated_length": 274.6875, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.0732, "grad_norm": 0.1582770049571991, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4285534.0, "reward": 0.8495199084281921, "reward_std": 0.6971250176429749, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5115375518798828, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 334.75, "completions/mean_terminated_length": 334.75, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.0734, "grad_norm": 0.081398606300354, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4294794.0, "reward": 0.42825740575790405, "reward_std": 0.4962455630302429, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16901248693466187, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 632.1875, "completions/mean_terminated_length": 632.1875, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.0736, "grad_norm": 0.041048552840948105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4309357.0, "reward": 0.41959837079048157, "reward_std": 0.37894269824028015, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08919374644756317, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 338.0, "completions/max_terminated_length": 338.0, "completions/mean_length": 238.75, "completions/mean_terminated_length": 238.75, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0738, "grad_norm": 0.0001438881445210427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4317401.0, "reward": 1.4993354082107544, "reward_std": 0.0006058369763195515, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986687898635864, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 383.875, "completions/mean_terminated_length": 383.875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.074, "grad_norm": 0.10590154677629471, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4327559.0, "reward": 0.6387355327606201, "reward_std": 0.559287965297699, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.277468740940094, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.966249942779541, "step": 370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 481.125, "completions/mean_terminated_length": 444.933349609375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.0742, "grad_norm": 0.0819612443447113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4339609.0, "reward": 0.4461601972579956, "reward_std": 0.6027922034263611, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2673187553882599, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.7224999666213989, "step": 371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 456.0625, "completions/mean_terminated_length": 456.0625, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.0744, "grad_norm": 0.09361572563648224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4350938.0, "reward": 0.634526252746582, "reward_std": 0.6434323787689209, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3315500020980835, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9756249785423279, "step": 372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 909.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 626.5625, "completions/mean_terminated_length": 626.5625, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0746, "grad_norm": 0.0746760368347168, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4365171.0, "reward": 0.6864850521087646, "reward_std": 0.6216354966163635, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.49796876311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.690625011920929, "step": 373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 435.875, "completions/mean_terminated_length": 435.875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.0748, "grad_norm": 0.08201146870851517, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4375937.0, "reward": 0.9779509902000427, "reward_std": 0.6368808150291443, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5809000134468079, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 615.5, "completions/mean_terminated_length": 615.5, "completions/min_length": 436.0, "completions/min_terminated_length": 436.0, "epoch": 0.075, "grad_norm": 0.07171034812927246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4390441.0, "reward": 0.8134857416152954, "reward_std": 0.7266150712966919, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5019687414169312, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 549.0, "completions/mean_terminated_length": 390.66668701171875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.0752, "grad_norm": 0.04268595948815346, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4402777.0, "reward": 0.336875855922699, "reward_std": 0.2699703276157379, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04874999821186066, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.71875, "step": 376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 407.1875, "completions/mean_terminated_length": 407.1875, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0754, "grad_norm": 0.055355701595544815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4414524.0, "reward": 0.4551886022090912, "reward_std": 0.36298295855522156, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09787499904632568, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 624.5625, "completions/mean_terminated_length": 624.5625, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.0756, "grad_norm": 0.07361561805009842, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4429285.0, "reward": 0.4517194926738739, "reward_std": 0.5348336100578308, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2784374952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.6937500238418579, "step": 378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 411.3125, "completions/mean_terminated_length": 411.3125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.0758, "grad_norm": 0.09500972926616669, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4439962.0, "reward": 0.8806948661804199, "reward_std": 0.6669890284538269, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5113875269889832, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 453.75, "completions/mean_terminated_length": 453.75, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.076, "grad_norm": 0.04586727172136307, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4451510.0, "reward": 0.5364855527877808, "reward_std": 0.31672489643096924, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1354687511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 547.0, "completions/mean_terminated_length": 547.0, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.0762, "grad_norm": 0.05021500959992409, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4464022.0, "reward": 0.5288041830062866, "reward_std": 0.36100971698760986, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18260625004768372, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7975000143051147, "step": 381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 415.625, "completions/mean_terminated_length": 415.625, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.0764, "grad_norm": 0.11359891295433044, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4474880.0, "reward": 0.7337385416030884, "reward_std": 0.6521686911582947, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4049749970436096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9137500524520874, "step": 382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 426.6875, "completions/mean_terminated_length": 426.6875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0766, "grad_norm": 0.077146515250206, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4485563.0, "reward": 0.925997793674469, "reward_std": 0.5341538190841675, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.60199373960495, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8068749904632568, "step": 383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 292.375, "completions/mean_terminated_length": 292.375, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.0768, "grad_norm": 0.13130244612693787, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4493953.0, "reward": 1.129807472229004, "reward_std": 0.5861271023750305, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.697112500667572, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.986875057220459, "step": 384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 637.75, "completions/mean_terminated_length": 612.0000610351562, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.077, "grad_norm": 0.06822662800550461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4508861.0, "reward": 0.5339852571487427, "reward_std": 0.5509573817253113, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2554687559604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8799999952316284, "step": 385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 446.75, "completions/mean_terminated_length": 446.75, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.0772, "grad_norm": 0.01772933639585972, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4519545.0, "reward": 0.5053635239601135, "reward_std": 0.13578540086746216, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0732249990105629, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 357.1875, "completions/mean_terminated_length": 357.1875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.0774, "grad_norm": 0.062404874712228775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4529292.0, "reward": 1.2868198156356812, "reward_std": 0.47135409712791443, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8236374855041504, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.96875, "step": 387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 298.875, "completions/mean_terminated_length": 298.875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.0776, "grad_norm": 0.11687123030424118, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4537914.0, "reward": 1.1648167371749878, "reward_std": 0.5270713567733765, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7046312093734741, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 404.75, "completions/mean_terminated_length": 363.4666748046875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0778, "grad_norm": 0.1427381932735443, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4548598.0, "reward": 0.7504574656486511, "reward_std": 0.7298926115036011, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5009124875068665, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7381250262260437, "step": 389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 391.5625, "completions/mean_terminated_length": 391.5625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.078, "grad_norm": 0.08677049726247787, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4558623.0, "reward": 0.3455919921398163, "reward_std": 0.5078069567680359, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1286812424659729, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9256249666213989, "step": 390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 445.375, "completions/mean_terminated_length": 445.375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.0782, "grad_norm": 0.0858520120382309, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4570533.0, "reward": 0.44523242115974426, "reward_std": 0.5068783164024353, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2029625028371811, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9337500333786011, "step": 391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 594.8125, "completions/mean_terminated_length": 594.8125, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.0784, "grad_norm": 0.05443073809146881, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4584562.0, "reward": 0.640066921710968, "reward_std": 0.5593204498291016, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2801312506198883, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.971875011920929, "step": 392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 283.875, "completions/mean_terminated_length": 283.875, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.0786, "grad_norm": 0.13965760171413422, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4592720.0, "reward": 0.7147542238235474, "reward_std": 0.5018692016601562, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30450624227523804, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 308.5, "completions/mean_terminated_length": 308.5, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.0788, "grad_norm": 0.14899100363254547, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4601848.0, "reward": 1.0345982313156128, "reward_std": 0.6413254737854004, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6316937208175659, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9818750619888306, "step": 394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 758.25, "completions/mean_terminated_length": 720.2857666015625, "completions/min_length": 487.0, "completions/min_terminated_length": 487.0, "epoch": 0.079, "grad_norm": 0.035836365073919296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4619084.0, "reward": 0.11343859136104584, "reward_std": 0.3116122782230377, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10187499970197678, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.7193750143051147, "step": 395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 300.0, "completions/max_terminated_length": 300.0, "completions/mean_length": 230.375, "completions/mean_terminated_length": 230.375, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.0792, "grad_norm": 0.19199275970458984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4626514.0, "reward": 1.1248390674591064, "reward_std": 0.6707237958908081, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7496750354766846, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9906250238418579, "step": 396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 607.1875, "completions/mean_terminated_length": 547.6428833007812, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0794, "grad_norm": 0.052567653357982635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4639861.0, "reward": 0.6622539758682251, "reward_std": 0.45524677634239197, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2620062232017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8268750309944153, "step": 397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 412.1875, "completions/mean_terminated_length": 412.1875, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.0796, "grad_norm": 0.050508901476860046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4651544.0, "reward": 0.2500016689300537, "reward_std": 0.2581992745399475, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 478.1875, "completions/mean_terminated_length": 400.21429443359375, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.0798, "grad_norm": 0.05250639468431473, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4663179.0, "reward": 0.3702355921268463, "reward_std": 0.31883710622787476, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11546874791383743, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.721875011920929, "step": 399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 317.75, "completions/mean_terminated_length": 317.75, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.08, "grad_norm": 0.13015252351760864, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4672135.0, "reward": 1.158644676208496, "reward_std": 0.6213160753250122, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7547875046730042, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.966249942779541, "step": 400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 516.375, "completions/mean_terminated_length": 516.375, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.0802, "grad_norm": 0.053035784512758255, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4683853.0, "reward": 0.3294261693954468, "reward_std": 0.4123680591583252, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09634999930858612, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 526.1875, "completions/mean_terminated_length": 526.1875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0804, "grad_norm": 0.06124286726117134, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4698384.0, "reward": 0.5739040374755859, "reward_std": 0.28234559297561646, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14780624210834503, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 884.0625, "completions/mean_terminated_length": 820.45458984375, "completions/min_length": 586.0, "completions/min_terminated_length": 586.0, "epoch": 0.0806, "grad_norm": 6.63916992493796e-08, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4716993.0, "reward": 1.1849999737023609e-06, "reward_std": 8.347854532075871e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.6225000023841858, "step": 403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 513.875, "completions/mean_terminated_length": 513.875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.0808, "grad_norm": 0.03900686651468277, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4729295.0, "reward": 0.5648446083068848, "reward_std": 0.24848397076129913, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25468748807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.778124988079071, "step": 404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 752.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 471.0, "completions/mean_terminated_length": 471.0, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.081, "grad_norm": 0.09138629585504532, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4740623.0, "reward": 0.5123445391654968, "reward_std": 0.40163761377334595, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3371874988079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6524999737739563, "step": 405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1011.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 702.25, "completions/mean_terminated_length": 702.25, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.0812, "grad_norm": 0.0631309375166893, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4756243.0, "reward": 0.5319418907165527, "reward_std": 0.5845411419868469, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3138812780380249, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7643749713897705, "step": 406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 331.625, "completions/mean_terminated_length": 331.625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0814, "grad_norm": 0.0808829739689827, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4765597.0, "reward": 1.080235481262207, "reward_std": 0.4905792474746704, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5979688167572021, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 478.4375, "completions/mean_terminated_length": 478.4375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.0816, "grad_norm": 0.09409550577402115, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4778212.0, "reward": 0.7368510961532593, "reward_std": 0.5692624449729919, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34869998693466187, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 262.9375, "completions/mean_terminated_length": 262.9375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.0818, "grad_norm": 0.10926062613725662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4786275.0, "reward": 0.6244509220123291, "reward_std": 0.4846191704273224, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24889999628067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 216.1875, "completions/mean_terminated_length": 216.1875, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 0.082, "grad_norm": 0.2662750482559204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4793574.0, "reward": 0.6977791786193848, "reward_std": 0.6771737933158875, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39555624127388, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8981250524520874, "step": 410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 494.75, "completions/mean_terminated_length": 494.75, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.0822, "grad_norm": 0.08958548307418823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4805026.0, "reward": 0.5735946893692017, "reward_std": 0.45496347546577454, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2721874713897705, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.768750011920929, "step": 411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 336.0, "completions/mean_terminated_length": 336.0, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.0824, "grad_norm": 0.09439997375011444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4814594.0, "reward": 0.40241044759750366, "reward_std": 0.39737510681152344, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11731874942779541, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 422.625, "completions/mean_terminated_length": 422.625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.0826, "grad_norm": 0.1073971837759018, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4825276.0, "reward": 0.514763355255127, "reward_std": 0.614412784576416, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2795250117778778, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.6618750095367432, "step": 413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 447.75, "completions/mean_terminated_length": 447.75, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0828, "grad_norm": 0.09176085889339447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4838024.0, "reward": 0.8691415190696716, "reward_std": 0.6065781116485596, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48828125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8468749523162842, "step": 414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 216.5, "completions/mean_terminated_length": 216.5, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.083, "grad_norm": 0.1358652114868164, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4845696.0, "reward": 1.2209291458129883, "reward_std": 0.5424522161483765, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8168562650680542, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9131249785423279, "step": 415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 495.4375, "completions/mean_terminated_length": 495.4375, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.0832, "grad_norm": 0.0841289684176445, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4857191.0, "reward": 0.6152320504188538, "reward_std": 0.35187453031539917, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2929624915122986, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8306249380111694, "step": 416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 285.375, "completions/mean_terminated_length": 285.375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.0834, "grad_norm": 0.19212430715560913, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4865245.0, "reward": 1.1228915452957153, "reward_std": 0.6695786118507385, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7457812428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 257.6875, "completions/mean_terminated_length": 257.6875, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.0836, "grad_norm": 0.17450402677059174, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4874008.0, "reward": 1.0965386629104614, "reward_std": 0.6351043581962585, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6930750012397766, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 403.25, "completions/mean_terminated_length": 403.25, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0838, "grad_norm": 0.10803667455911636, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4884316.0, "reward": 0.6761698722839355, "reward_std": 0.5307461023330688, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28983747959136963, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 435.0, "completions/mean_terminated_length": 395.7333679199219, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.084, "grad_norm": 0.12259797751903534, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4894924.0, "reward": 0.9804542660713196, "reward_std": 0.6356059908866882, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5859062671661377, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9043750166893005, "step": 420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 423.5, "completions/mean_terminated_length": 423.5, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.0842, "grad_norm": 0.036517709493637085, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4905556.0, "reward": 0.39519184827804565, "reward_std": 0.23629014194011688, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04038124904036522, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9756249785423279, "step": 421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 339.0, "completions/max_terminated_length": 339.0, "completions/mean_length": 264.625, "completions/mean_terminated_length": 264.625, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0844, "grad_norm": 0.17559713125228882, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4913950.0, "reward": 0.7871826887130737, "reward_std": 0.6797050833702087, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.44936245679855347, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 715.4375, "completions/mean_terminated_length": 612.5833740234375, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0846, "grad_norm": 0.09370304644107819, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4929925.0, "reward": 0.44031351804733276, "reward_std": 0.6551275849342346, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.25562500953674316, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 721.5, "completions/mean_terminated_length": 678.2857666015625, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.0848, "grad_norm": 0.09862291812896729, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4946029.0, "reward": 0.4678916931152344, "reward_std": 0.7167544960975647, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31078124046325684, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.828125, "step": 424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 569.8125, "completions/mean_terminated_length": 569.8125, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.085, "grad_norm": 0.04315030574798584, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4959946.0, "reward": 0.4599137306213379, "reward_std": 0.3647184371948242, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1073250025510788, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.96875, "step": 425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 723.4375, "completions/mean_terminated_length": 703.4000244140625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.0852, "grad_norm": 0.02319013513624668, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4975297.0, "reward": 0.4370200037956238, "reward_std": 0.221146360039711, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.061537500470876694, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 362.625, "completions/mean_terminated_length": 362.625, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.0854, "grad_norm": 0.10660771280527115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4985195.0, "reward": 0.94648277759552, "reward_std": 0.5951193571090698, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5179624557495117, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 282.0, "completions/max_terminated_length": 282.0, "completions/mean_length": 240.25, "completions/mean_terminated_length": 240.25, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.0856, "grad_norm": 0.04692617431282997, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4992591.0, "reward": 0.45708245038986206, "reward_std": 0.17949891090393066, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03916250169277191, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 433.375, "completions/mean_terminated_length": 433.375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0858, "grad_norm": 0.07389457523822784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5003749.0, "reward": 0.5158576965332031, "reward_std": 0.44780662655830383, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15671250224113464, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 155.125, "completions/mean_terminated_length": 155.125, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 0.086, "grad_norm": 0.006451352033764124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5009783.0, "reward": 0.5158262252807617, "reward_std": 0.02378176897764206, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03164999932050705, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 489.0, "completions/max_terminated_length": 489.0, "completions/mean_length": 316.25, "completions/mean_terminated_length": 316.25, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0862, "grad_norm": 0.08249016106128693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5018363.0, "reward": 0.845822811126709, "reward_std": 0.4550795257091522, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37914374470710754, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 163.0, "completions/max_terminated_length": 163.0, "completions/mean_length": 132.3125, "completions/mean_terminated_length": 132.3125, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "epoch": 0.0864, "grad_norm": 0.26224783062934875, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5024256.0, "reward": 0.5943738222122192, "reward_std": 0.49041488766670227, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1887437403202057, "rewards/confidence_one_or_zero": 0.9375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9993749856948853, "step": 432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 478.375, "completions/mean_terminated_length": 478.375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.0866, "grad_norm": 0.08558433502912521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5035798.0, "reward": 0.955470085144043, "reward_std": 0.5884618759155273, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.535937488079071, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 430.25, "completions/mean_terminated_length": 430.25, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.0868, "grad_norm": 0.07268156856298447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5046970.0, "reward": 0.6816259622573853, "reward_std": 0.5312881469726562, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3007500171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 469.375, "completions/mean_terminated_length": 432.4000244140625, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.087, "grad_norm": 0.054923586547374725, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5058560.0, "reward": 0.6492915153503418, "reward_std": 0.33070725202560425, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.29858124256134033, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7331249713897705, "step": 435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 197.0, "completions/max_terminated_length": 197.0, "completions/mean_length": 151.1875, "completions/mean_terminated_length": 151.1875, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.0872, "grad_norm": 0.14707203209400177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5064595.0, "reward": 1.344282865524292, "reward_std": 0.435295045375824, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8760624527931213, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9937500357627869, "step": 436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 307.0, "completions/max_terminated_length": 307.0, "completions/mean_length": 226.0625, "completions/mean_terminated_length": 226.0625, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.0874, "grad_norm": 0.12090739607810974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5071828.0, "reward": 0.7479447722434998, "reward_std": 0.46800729632377625, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30838748812675476, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 423.0625, "completions/mean_terminated_length": 423.0625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.0876, "grad_norm": 0.04039503633975983, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5082053.0, "reward": 0.5141289830207825, "reward_std": 0.21658189594745636, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15325623750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8737499713897705, "step": 438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 608.6875, "completions/mean_terminated_length": 608.6875, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.0878, "grad_norm": 0.04654844105243683, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5096128.0, "reward": 0.2661728858947754, "reward_std": 0.3237794041633606, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09484374523162842, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.699999988079071, "step": 439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 719.125, "completions/mean_terminated_length": 617.5, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.088, "grad_norm": 0.0601293221116066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5111506.0, "reward": 0.3341163694858551, "reward_std": 0.4675639867782593, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.10573124885559082, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.5631250143051147, "step": 440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 732.5625, "completions/mean_terminated_length": 713.1333618164062, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.0882, "grad_norm": 0.048681799322366714, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5126971.0, "reward": 0.5224226713180542, "reward_std": 0.3923233449459076, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23234376311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 598.375, "completions/mean_terminated_length": 598.375, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.0884, "grad_norm": 0.041828956454992294, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5140801.0, "reward": 0.32945114374160767, "reward_std": 0.3181913197040558, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09639999270439148, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8212499618530273, "step": 442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 360.375, "completions/mean_terminated_length": 360.375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.0886, "grad_norm": 0.11021311581134796, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5150247.0, "reward": 1.0694166421890259, "reward_std": 0.5290600657463074, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6388312578201294, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 251.8125, "completions/mean_terminated_length": 251.8125, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.0888, "grad_norm": 0.11398208141326904, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5157908.0, "reward": 1.3241229057312012, "reward_std": 0.3771698772907257, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.83574378490448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 350.5, "completions/mean_terminated_length": 350.5, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.089, "grad_norm": 0.08058730512857437, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5168156.0, "reward": 1.2270324230194092, "reward_std": 0.5014961957931519, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7665624618530273, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 614.1875, "completions/mean_terminated_length": 614.1875, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.0892, "grad_norm": 0.05954090505838394, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5182143.0, "reward": 0.3728010952472687, "reward_std": 0.36808231472969055, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18310000002384186, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7831250429153442, "step": 446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 395.5625, "completions/mean_terminated_length": 395.5625, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.0894, "grad_norm": 0.06562965363264084, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5192392.0, "reward": 0.5636415481567383, "reward_std": 0.42573675513267517, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18978124856948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 175.0, "completions/max_terminated_length": 175.0, "completions/mean_length": 148.375, "completions/mean_terminated_length": 148.375, "completions/min_length": 118.0, "completions/min_terminated_length": 118.0, "epoch": 0.0896, "grad_norm": 0.07480353116989136, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5198478.0, "reward": 1.4377011060714722, "reward_std": 0.24740071594715118, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9378999471664429, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 614.8125, "completions/mean_terminated_length": 614.8125, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.0898, "grad_norm": 0.055710144340991974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5212203.0, "reward": 0.36850085854530334, "reward_std": 0.35339120030403137, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17449992895126343, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8280624747276306, "step": 449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 321.3125, "completions/mean_terminated_length": 321.3125, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.09, "grad_norm": 0.13426007330417633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5221728.0, "reward": 0.8212102651596069, "reward_std": 0.6500993967056274, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.454918771982193, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 419.0625, "completions/mean_terminated_length": 419.0625, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.0902, "grad_norm": 0.10001935809850693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5232049.0, "reward": 0.7044854760169983, "reward_std": 0.5057511329650879, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2839687466621399, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9737500548362732, "step": 451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 149.3125, "completions/mean_terminated_length": 149.3125, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.0904, "grad_norm": 0.005847614724189043, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5237766.0, "reward": 0.5116919875144958, "reward_std": 0.015601950697600842, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.023381248116493225, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9881187677383423, "step": 452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 257.125, "completions/mean_terminated_length": 257.125, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.0906, "grad_norm": 0.11528044939041138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5245736.0, "reward": 1.1629482507705688, "reward_std": 0.4634849429130554, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.700893759727478, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9268749952316284, "step": 453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 475.6875, "completions/mean_terminated_length": 475.6875, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.0908, "grad_norm": 0.06797628104686737, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5257459.0, "reward": 0.7545322179794312, "reward_std": 0.4650496542453766, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3215624988079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 310.3125, "completions/mean_terminated_length": 310.3125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.091, "grad_norm": 0.050952911376953125, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5265960.0, "reward": 0.5738290548324585, "reward_std": 0.2820831835269928, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 378.125, "completions/mean_terminated_length": 378.125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.0912, "grad_norm": 0.08675990998744965, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5276042.0, "reward": 0.834060549736023, "reward_std": 0.4632284939289093, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35561877489089966, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 437.6875, "completions/mean_terminated_length": 437.6875, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.0914, "grad_norm": 0.03535059466958046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5287077.0, "reward": 0.4945946931838989, "reward_std": 0.20074652135372162, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11418749392032623, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9262499809265137, "step": 457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 726.5, "completions/mean_terminated_length": 657.84619140625, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.0916, "grad_norm": 0.0453319177031517, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5303309.0, "reward": 0.45250070095062256, "reward_std": 0.4025155305862427, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2174999862909317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 249.25, "completions/mean_terminated_length": 249.25, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.0918, "grad_norm": 0.041174717247486115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5311329.0, "reward": 0.48709505796432495, "reward_std": 0.13158337771892548, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.036687497049570084, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 321.5, "completions/mean_terminated_length": 321.5, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.092, "grad_norm": 0.08916441351175308, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5320281.0, "reward": 1.315591812133789, "reward_std": 0.3950507342815399, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8186812400817871, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706249833106995, "step": 460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 296.0625, "completions/mean_terminated_length": 296.0625, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0922, "grad_norm": 0.06554467231035233, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5328794.0, "reward": 0.6000008583068848, "reward_std": 0.2912408411502838, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19999998807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8562500476837158, "step": 461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 409.875, "completions/mean_terminated_length": 409.875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.0924, "grad_norm": 0.09337291121482849, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5339128.0, "reward": 0.7405418157577515, "reward_std": 0.471584677696228, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2935812473297119, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 740.4375, "completions/mean_terminated_length": 675.0, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.0926, "grad_norm": 0.059284795075654984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5354991.0, "reward": 0.7467974424362183, "reward_std": 0.5443063974380493, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6810937523841858, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.21437501907348633, "step": 463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 378.0, "completions/mean_terminated_length": 378.0, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.0928, "grad_norm": 0.023945312947034836, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5365391.0, "reward": 0.5066009759902954, "reward_std": 0.13582350313663483, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07569999992847443, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 436.0, "completions/mean_length": 328.125, "completions/mean_terminated_length": 281.73333740234375, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.093, "grad_norm": 0.15464499592781067, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5374257.0, "reward": 0.8488917350769043, "reward_std": 0.6214277148246765, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4477812349796295, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 952.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 528.375, "completions/mean_terminated_length": 528.375, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.0932, "grad_norm": 0.08950545638799667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5387239.0, "reward": 0.6748415231704712, "reward_std": 0.56363445520401, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3496812582015991, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8399999737739563, "step": 466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 704.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 527.125, "completions/mean_terminated_length": 527.125, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0934, "grad_norm": 0.09403914958238602, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5399881.0, "reward": 0.6977108716964722, "reward_std": 0.5401663780212402, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3329187333583832, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9193750023841858, "step": 467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 472.0, "completions/mean_terminated_length": 472.0, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.0936, "grad_norm": 0.05346408113837242, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5411401.0, "reward": 0.5972665548324585, "reward_std": 0.344707190990448, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.25703126192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.846875011920929, "step": 468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 386.3125, "completions/mean_terminated_length": 386.3125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.0938, "grad_norm": 0.039824843406677246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5421182.0, "reward": 0.46803227066993713, "reward_std": 0.18361663818359375, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06106249988079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9668687582015991, "step": 469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 652.4375, "completions/mean_terminated_length": 599.357177734375, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.094, "grad_norm": 0.06563054025173187, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5435461.0, "reward": 0.35162290930747986, "reward_std": 0.5099273920059204, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14074374735355377, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8474999666213989, "step": 470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 868.8125, "completions/mean_terminated_length": 775.7000122070312, "completions/min_length": 617.0, "completions/min_terminated_length": 617.0, "epoch": 0.0942, "grad_norm": 0.07232042402029037, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5453602.0, "reward": 0.3668754994869232, "reward_std": 0.5441860556602478, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23375001549720764, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.4468749761581421, "step": 471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 282.0, "completions/max_terminated_length": 282.0, "completions/mean_length": 227.25, "completions/mean_terminated_length": 227.25, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.0944, "grad_norm": 0.0002268613752676174, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5462102.0, "reward": 1.4987385272979736, "reward_std": 0.0011043275007978082, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9974749684333801, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 794.625, "completions/mean_terminated_length": 779.3333740234375, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.0946, "grad_norm": 0.0733872726559639, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5479024.0, "reward": 0.7612477540969849, "reward_std": 0.6505748629570007, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.45999374985694885, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8400000333786011, "step": 473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 390.6875, "completions/mean_terminated_length": 348.4666748046875, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.0948, "grad_norm": 0.1515096127986908, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5488907.0, "reward": 1.0412228107452393, "reward_std": 0.5512046217918396, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5824437737464905, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9081250429153442, "step": 474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 414.5, "completions/mean_terminated_length": 414.5, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.095, "grad_norm": 0.10122788697481155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5499571.0, "reward": 1.0341415405273438, "reward_std": 0.578716516494751, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6307812929153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.921875, "step": 475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 557.875, "completions/mean_terminated_length": 491.2857360839844, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.0952, "grad_norm": 0.08874009549617767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5512657.0, "reward": 0.7025009393692017, "reward_std": 0.6244086623191833, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4049999713897705, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.71875, "step": 476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 457.375, "completions/mean_terminated_length": 457.375, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.0954, "grad_norm": 0.04316806048154831, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5523895.0, "reward": 0.5138821601867676, "reward_std": 0.2257566601037979, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1527625024318695, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 388.75, "completions/mean_terminated_length": 388.75, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0956, "grad_norm": 0.10690197348594666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5533763.0, "reward": 1.037029504776001, "reward_std": 0.6392847895622253, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6365562677383423, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9275000095367432, "step": 478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 367.25, "completions/mean_terminated_length": 367.25, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.0958, "grad_norm": 0.10267863422632217, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5543527.0, "reward": 0.8911573886871338, "reward_std": 0.5804669857025146, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4698125123977661, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 584.125, "completions/mean_terminated_length": 554.800048828125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.096, "grad_norm": 0.058116428554058075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5556441.0, "reward": 0.5568321943283081, "reward_std": 0.42533406615257263, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1761625111103058, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9043749570846558, "step": 480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 440.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 273.0, "completions/mean_terminated_length": 273.0, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.0962, "grad_norm": 0.08120523393154144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5564921.0, "reward": 0.7198199033737183, "reward_std": 0.3916865587234497, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3771374821662903, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7693749666213989, "step": 481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 296.75, "completions/mean_terminated_length": 296.75, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.0964, "grad_norm": 0.12573523819446564, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5573621.0, "reward": 0.8992040753364563, "reward_std": 0.5751785039901733, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4859062433242798, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 411.0, "completions/mean_terminated_length": 411.0, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.0966, "grad_norm": 0.11784812062978745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5584469.0, "reward": 0.9736577272415161, "reward_std": 0.6409822702407837, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5723124742507935, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9925000071525574, "step": 483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 395.0625, "completions/mean_terminated_length": 395.0625, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.0968, "grad_norm": 0.10219369828701019, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5595270.0, "reward": 0.6147665977478027, "reward_std": 0.5799944400787354, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2920312285423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.949999988079071, "step": 484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 259.0, "completions/max_terminated_length": 259.0, "completions/mean_length": 198.0625, "completions/mean_terminated_length": 198.0625, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.097, "grad_norm": 0.1163184866309166, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5602247.0, "reward": 1.4057354927062988, "reward_std": 0.37486299872398376, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9364687204360962, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 793.625, "completions/mean_terminated_length": 760.7142944335938, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.0972, "grad_norm": 0.0512387789785862, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5619201.0, "reward": 0.4750661849975586, "reward_std": 0.4103899598121643, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32513123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5581250190734863, "step": 486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 352.5625, "completions/mean_terminated_length": 352.5625, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.0974, "grad_norm": 0.09402058273553848, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5628778.0, "reward": 0.8259233236312866, "reward_std": 0.4697376787662506, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33934372663497925, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 603.0625, "completions/mean_terminated_length": 603.0625, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.0976, "grad_norm": 0.03807405009865761, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5642459.0, "reward": 0.38913846015930176, "reward_std": 0.2841024398803711, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0907750055193901, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8637499809265137, "step": 488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 590.0, "completions/mean_terminated_length": 561.0667114257812, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.0978, "grad_norm": 0.06939234584569931, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5656091.0, "reward": 0.9500790238380432, "reward_std": 0.5516497492790222, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5876562595367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.78125, "step": 489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 467.5625, "completions/mean_terminated_length": 467.5625, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.098, "grad_norm": 0.014199239201843739, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5667620.0, "reward": 0.5741415023803711, "reward_std": 0.10156233608722687, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1482812464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 347.0, "completions/mean_terminated_length": 347.0, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.0982, "grad_norm": 0.105892114341259, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5677572.0, "reward": 0.8355792164802551, "reward_std": 0.46273279190063477, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35865625739097595, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 426.6875, "completions/mean_terminated_length": 426.6875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0984, "grad_norm": 0.033437494188547134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5687951.0, "reward": 0.4650229811668396, "reward_std": 0.18235965073108673, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.055043745785951614, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9093750715255737, "step": 492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.0986, "grad_norm": 0.15008582174777985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5697009.0, "reward": 1.1956260204315186, "reward_std": 0.4948778748512268, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7662500143051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9124999642372131, "step": 493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 370.125, "completions/mean_terminated_length": 370.125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.0988, "grad_norm": 0.09588021039962769, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5706915.0, "reward": 0.8884631395339966, "reward_std": 0.5189713835716248, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4644243121147156, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9320833683013916, "step": 494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 544.375, "completions/mean_terminated_length": 544.375, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.099, "grad_norm": 0.07364633679389954, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5719465.0, "reward": 0.6315605640411377, "reward_std": 0.48664408922195435, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2631187438964844, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9493749737739563, "step": 495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 378.9375, "completions/mean_terminated_length": 378.9375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0992, "grad_norm": 0.11381593346595764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5729496.0, "reward": 0.7114824056625366, "reward_std": 0.5027160048484802, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.297962486743927, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 980.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 603.125, "completions/mean_terminated_length": 603.125, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.0994, "grad_norm": 0.050265129655599594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5743738.0, "reward": 0.5376571416854858, "reward_std": 0.31590554118156433, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2628124952316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7000000476837158, "step": 497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 392.5625, "completions/mean_terminated_length": 392.5625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.0996, "grad_norm": 0.029865749180316925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5753763.0, "reward": 0.49931979179382324, "reward_std": 0.1343555748462677, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.061137497425079346, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 840.0, "completions/mean_length": 552.625, "completions/mean_terminated_length": 521.2000122070312, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0998, "grad_norm": 0.060566700994968414, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5766397.0, "reward": 0.6154696345329285, "reward_std": 0.4019663333892822, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35593748092651367, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.703125, "step": 499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 533.5, "completions/mean_terminated_length": 533.5, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.1, "grad_norm": 0.05195105820894241, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5779461.0, "reward": 0.5657791495323181, "reward_std": 0.32940971851348877, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19405624270439148, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 463.0625, "completions/mean_terminated_length": 425.66668701171875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.1002, "grad_norm": 0.08610744029283524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5790998.0, "reward": 0.7441259026527405, "reward_std": 0.5038689374923706, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3632500171661377, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8356250524520874, "step": 501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 439.4375, "completions/mean_terminated_length": 439.4375, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.1004, "grad_norm": 0.00336195295676589, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5802029.0, "reward": 0.5323792695999146, "reward_std": 0.01998257078230381, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06475624442100525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 291.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 231.5625, "completions/mean_terminated_length": 231.5625, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.1006, "grad_norm": 0.14057335257530212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5810166.0, "reward": 1.0428760051727295, "reward_std": 0.5493088364601135, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5857499837875366, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 252.0, "completions/max_terminated_length": 252.0, "completions/mean_length": 181.375, "completions/mean_terminated_length": 181.375, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 0.1008, "grad_norm": 0.14156264066696167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5816700.0, "reward": 0.8826919794082642, "reward_std": 0.4935321807861328, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3903812766075134, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 457.8125, "completions/mean_terminated_length": 457.8125, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.101, "grad_norm": 0.043507304042577744, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5828041.0, "reward": 0.5996073484420776, "reward_std": 0.24046847224235535, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13671249151229858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 633.6875, "completions/mean_terminated_length": 607.6666870117188, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.1012, "grad_norm": 0.021834038197994232, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5842436.0, "reward": 0.4859384000301361, "reward_std": 0.1903340220451355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09687499701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 302.25, "completions/mean_terminated_length": 302.25, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.1014, "grad_norm": 0.10129579901695251, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5851208.0, "reward": 0.9053131341934204, "reward_std": 0.5216514468193054, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6231250166893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49687498807907104, "step": 507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 418.25, "completions/mean_terminated_length": 418.25, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1016, "grad_norm": 0.08457519859075546, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5861852.0, "reward": 0.6616164445877075, "reward_std": 0.41414543986320496, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3232312798500061, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8100000023841858, "step": 508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 290.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 226.125, "completions/mean_terminated_length": 226.125, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.1018, "grad_norm": 0.15134869515895844, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5869470.0, "reward": 0.8619760870933533, "reward_std": 0.5268114805221558, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41145002841949463, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9718750715255737, "step": 509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 337.625, "completions/mean_terminated_length": 337.625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.102, "grad_norm": 0.0021516389679163694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5878808.0, "reward": 0.5445165634155273, "reward_std": 0.011709420010447502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 240.9375, "completions/mean_terminated_length": 240.9375, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.1022, "grad_norm": 0.19083434343338013, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5886423.0, "reward": 1.133507251739502, "reward_std": 0.5811230540275574, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7045125365257263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 444.4375, "completions/mean_terminated_length": 444.4375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.1024, "grad_norm": 0.12572282552719116, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5897358.0, "reward": 0.9864072203636169, "reward_std": 0.5452103018760681, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5353124737739563, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 364.875, "completions/mean_terminated_length": 364.875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.1026, "grad_norm": 0.09817007184028625, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5907196.0, "reward": 1.200207233428955, "reward_std": 0.45769649744033813, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7129125595092773, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 664.5, "completions/mean_terminated_length": 640.5333862304688, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.1028, "grad_norm": 0.06430813670158386, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5922644.0, "reward": 0.4794541895389557, "reward_std": 0.45004063844680786, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.27140623331069946, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.675000011920929, "step": 514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 399.0, "completions/max_terminated_length": 399.0, "completions/mean_length": 319.4375, "completions/mean_terminated_length": 319.4375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.103, "grad_norm": 0.1121593564748764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5931435.0, "reward": 1.1642323732376099, "reward_std": 0.5283159613609314, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7034624814987183, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 393.0625, "completions/mean_terminated_length": 393.0625, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.1032, "grad_norm": 0.0816410481929779, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5942396.0, "reward": 1.1377352476119995, "reward_std": 0.5007569789886475, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7129687666893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.893750011920929, "step": 516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 404.9375, "completions/mean_terminated_length": 404.9375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.1034, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5952939.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 655.125, "completions/mean_terminated_length": 570.0, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1036, "grad_norm": 0.10155541449785233, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5967501.0, "reward": 0.7257789373397827, "reward_std": 0.6487346291542053, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5140562057495117, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5525000095367432, "step": 518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 299.3125, "completions/mean_terminated_length": 299.3125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.1038, "grad_norm": 0.13066017627716064, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5976594.0, "reward": 1.1012449264526367, "reward_std": 0.5739299058914185, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7024874687194824, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8612499833106995, "step": 519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 275.5625, "completions/mean_terminated_length": 275.5625, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.104, "grad_norm": 0.14879567921161652, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5984827.0, "reward": 0.9507794380187988, "reward_std": 0.4588189423084259, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5265562534332275, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8806250095367432, "step": 520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 406.0625, "completions/mean_terminated_length": 406.0625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.1042, "grad_norm": 0.07158734649419785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5995164.0, "reward": 0.5871853232383728, "reward_std": 0.28591078519821167, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.174368754029274, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9356250166893005, "step": 521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 582.6875, "completions/mean_terminated_length": 582.6875, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.1044, "grad_norm": 0.06702210754156113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6009895.0, "reward": 0.930469274520874, "reward_std": 0.38185936212539673, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6734374761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.48124998807907104, "step": 522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 498.3125, "completions/mean_terminated_length": 498.3125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.1046, "grad_norm": 0.06123201921582222, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6021948.0, "reward": 0.5197665095329285, "reward_std": 0.36385056376457214, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.28953123092651367, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 601.75, "completions/mean_terminated_length": 541.4285888671875, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.1048, "grad_norm": 0.07760372757911682, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6035480.0, "reward": 0.6350789070129395, "reward_std": 0.5543599724769592, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.39515626430511475, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.503125011920929, "step": 524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 231.875, "completions/mean_terminated_length": 231.875, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.105, "grad_norm": 0.14564433693885803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6042726.0, "reward": 0.6131795048713684, "reward_std": 0.3702974021434784, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16385625302791595, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9131250381469727, "step": 525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 542.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 360.375, "completions/mean_terminated_length": 360.375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.1052, "grad_norm": 0.06868156045675278, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6052764.0, "reward": 0.6283478736877441, "reward_std": 0.3658752143383026, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1941937506198883, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 387.0, "completions/max_terminated_length": 387.0, "completions/mean_length": 299.8125, "completions/mean_terminated_length": 299.8125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1054, "grad_norm": 0.10128405690193176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6061609.0, "reward": 1.0815417766571045, "reward_std": 0.4897078275680542, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6005812287330627, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 709.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 449.25, "completions/mean_terminated_length": 449.25, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.1056, "grad_norm": 0.035024918615818024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6073133.0, "reward": 0.509285569190979, "reward_std": 0.13867270946502686, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0810687467455864, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 703.4375, "completions/mean_terminated_length": 703.4375, "completions/min_length": 513.0, "completions/min_terminated_length": 513.0, "epoch": 0.1058, "grad_norm": 0.03593743219971657, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6088388.0, "reward": 0.4487510025501251, "reward_std": 0.22293764352798462, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08500000089406967, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9375, "step": 529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 479.5, "completions/mean_terminated_length": 479.5, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.106, "grad_norm": 0.03100421093404293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6100860.0, "reward": 0.5654697418212891, "reward_std": 0.19170920550823212, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19343750178813934, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 447.875, "completions/mean_terminated_length": 447.875, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1062, "grad_norm": 0.03574932739138603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6114186.0, "reward": 0.5813289880752563, "reward_std": 0.26004889607429504, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2876562476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 354.25, "completions/mean_terminated_length": 354.25, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.1064, "grad_norm": 0.0857049897313118, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6123662.0, "reward": 0.7269415855407715, "reward_std": 0.3830983638763428, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26638126373291016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 537.625, "completions/mean_terminated_length": 537.625, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.1066, "grad_norm": 0.04765480011701584, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6136792.0, "reward": 0.47931987047195435, "reward_std": 0.2762911021709442, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14613750576972961, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 706.5625, "completions/mean_terminated_length": 661.2142944335938, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.1068, "grad_norm": 0.0472337082028389, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6153297.0, "reward": 0.5497633218765259, "reward_std": 0.38222646713256836, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3495250344276428, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5900000333786011, "step": 534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 810.375, "completions/mean_terminated_length": 761.0769653320312, "completions/min_length": 570.0, "completions/min_terminated_length": 570.0, "epoch": 0.107, "grad_norm": 0.07172301411628723, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6170023.0, "reward": 0.6607789993286133, "reward_std": 0.6445109844207764, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3840562403202057, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7524999976158142, "step": 535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 507.6875, "completions/mean_terminated_length": 507.6875, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.1072, "grad_norm": 0.0033076368272304535, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6181698.0, "reward": 0.5396665930747986, "reward_std": 0.016399169340729713, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07933124899864197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 468.3125, "completions/mean_terminated_length": 431.2666931152344, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1074, "grad_norm": 0.044996749609708786, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6192951.0, "reward": 0.47921672463417053, "reward_std": 0.1936369091272354, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08343124389648438, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 320.0, "completions/mean_terminated_length": 320.0, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1076, "grad_norm": 0.11640621721744537, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6201767.0, "reward": 0.7718416452407837, "reward_std": 0.5373316407203674, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.356181263923645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 552.625, "completions/mean_terminated_length": 552.625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1078, "grad_norm": 0.06362716853618622, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6214481.0, "reward": 0.5932820439338684, "reward_std": 0.3936975598335266, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.31156250834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 741.9375, "completions/mean_terminated_length": 723.1333618164062, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.108, "grad_norm": 0.059128887951374054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6230224.0, "reward": 0.4737507700920105, "reward_std": 0.48418688774108887, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7112500071525574, "step": 540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 321.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 248.125, "completions/mean_terminated_length": 248.125, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.1082, "grad_norm": 0.12582120299339294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6238290.0, "reward": 0.9613291025161743, "reward_std": 0.48964911699295044, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48515626788139343, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 438.1875, "completions/mean_terminated_length": 438.1875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.1084, "grad_norm": 0.01633545011281967, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6248885.0, "reward": 0.5824227333068848, "reward_std": 0.11113829165697098, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16484373807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 216.9375, "completions/mean_terminated_length": 216.9375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.1086, "grad_norm": 0.1419299691915512, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6256468.0, "reward": 1.2676446437835693, "reward_std": 0.41465917229652405, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7852874994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 683.75, "completions/mean_terminated_length": 661.0667114257812, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.1088, "grad_norm": 0.05970469117164612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6271232.0, "reward": 0.673985481262207, "reward_std": 0.4667232632637024, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2854687571525574, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8518749475479126, "step": 544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 669.125, "completions/mean_terminated_length": 618.4285888671875, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.109, "grad_norm": 0.052359070628881454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6286754.0, "reward": 0.561641275882721, "reward_std": 0.36732935905456543, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.37328124046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.543749988079071, "step": 545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 438.8125, "completions/mean_terminated_length": 438.8125, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.1092, "grad_norm": 0.10258364677429199, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6298399.0, "reward": 0.5939667224884033, "reward_std": 0.5140591263771057, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25043123960494995, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 458.0, "completions/mean_terminated_length": 458.0, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.1094, "grad_norm": 0.09186069667339325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6310591.0, "reward": 0.8958603143692017, "reward_std": 0.5770253539085388, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4792187809944153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 431.9375, "completions/mean_terminated_length": 431.9375, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.1096, "grad_norm": 0.12965483963489532, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6321742.0, "reward": 0.9417044520378113, "reward_std": 0.6771584749221802, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5709062218666077, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 238.0, "completions/max_terminated_length": 238.0, "completions/mean_length": 172.6875, "completions/mean_terminated_length": 172.6875, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.1098, "grad_norm": 8.608058124082163e-06, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6327993.0, "reward": 1.4999828338623047, "reward_std": 2.545519055274781e-05, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9999624490737915, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9962500333786011, "step": 549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 799.5, "completions/mean_terminated_length": 767.4285888671875, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.11, "grad_norm": 0.051411233842372894, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6345329.0, "reward": 0.6536725759506226, "reward_std": 0.46567055583000183, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6198437213897705, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.27125000953674316, "step": 550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 426.875, "completions/mean_terminated_length": 426.875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.1102, "grad_norm": 0.08002384006977081, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6355791.0, "reward": 0.9614041447639465, "reward_std": 0.48973703384399414, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4853062331676483, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 330.9375, "completions/mean_terminated_length": 330.9375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.1104, "grad_norm": 0.08009073138237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6365214.0, "reward": 1.3192635774612427, "reward_std": 0.38711756467819214, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8260250091552734, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 631.8125, "completions/mean_terminated_length": 541.3077392578125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1106, "grad_norm": 0.05086054652929306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6379419.0, "reward": 0.374525785446167, "reward_std": 0.3209315240383148, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12404999881982803, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5299999713897705, "step": 553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 519.9375, "completions/mean_terminated_length": 447.9285888671875, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.1108, "grad_norm": 0.08195751160383224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6391754.0, "reward": 0.6148315072059631, "reward_std": 0.5200648307800293, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29216092824935913, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7359374761581421, "step": 554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 696.9375, "completions/mean_terminated_length": 650.2142944335938, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.111, "grad_norm": 0.04585365206003189, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6407145.0, "reward": 0.4630354046821594, "reward_std": 0.36469563841819763, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1135687455534935, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 467.625, "completions/mean_terminated_length": 467.625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1112, "grad_norm": 0.04577700421214104, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6418339.0, "reward": 0.6361728310585022, "reward_std": 0.3640844523906708, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20984375476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 596.5625, "completions/mean_terminated_length": 596.5625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.1114, "grad_norm": 0.02608029544353485, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6431820.0, "reward": 0.5325793623924255, "reward_std": 0.16644898056983948, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1276562511920929, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 789.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 487.375, "completions/mean_terminated_length": 487.375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.1116, "grad_norm": 0.08064794540405273, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6443682.0, "reward": 1.0723261833190918, "reward_std": 0.49992895126342773, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5821500420570374, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 468.4375, "completions/mean_terminated_length": 468.4375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1118, "grad_norm": 0.08925466239452362, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6455801.0, "reward": 0.781082034111023, "reward_std": 0.46206799149513245, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.49966248869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6862500309944153, "step": 559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 435.6875, "completions/mean_terminated_length": 435.6875, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.112, "grad_norm": 0.09641657024621964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6466900.0, "reward": 0.9846728444099426, "reward_std": 0.5468806028366089, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5318437814712524, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 475.75, "completions/mean_terminated_length": 475.75, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.1122, "grad_norm": 0.08427732437849045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6478768.0, "reward": 0.7826571464538574, "reward_std": 0.5321314930915833, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3778125047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 575.875, "completions/mean_terminated_length": 575.875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.1124, "grad_norm": 0.029340598732233047, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6492782.0, "reward": 0.6072664260864258, "reward_std": 0.194798544049263, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2770312428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 746.5625, "completions/mean_terminated_length": 530.7777709960938, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.1126, "grad_norm": 0.09430040419101715, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6509239.0, "reward": 0.4416411519050598, "reward_std": 0.652595579624176, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.25828123092651367, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5218750238418579, "step": 563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 382.3125, "completions/mean_terminated_length": 382.3125, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.1128, "grad_norm": 0.060440145432949066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6519372.0, "reward": 0.5647447109222412, "reward_std": 0.2834181785583496, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12948749959468842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 325.875, "completions/mean_terminated_length": 325.875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.113, "grad_norm": 0.09266310185194016, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6528122.0, "reward": 0.6764853000640869, "reward_std": 0.4552561640739441, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2904687523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 442.625, "completions/mean_terminated_length": 442.625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1132, "grad_norm": 0.0714007019996643, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6539252.0, "reward": 0.7646883726119995, "reward_std": 0.42428210377693176, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5293749570846558, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.668749988079071, "step": 566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 461.6875, "completions/mean_terminated_length": 461.6875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.1134, "grad_norm": 0.04012882336974144, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6550783.0, "reward": 0.6541382670402527, "reward_std": 0.2529754340648651, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3707750141620636, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6712499856948853, "step": 567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 282.375, "completions/mean_terminated_length": 282.375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.1136, "grad_norm": 0.11355920135974884, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6559269.0, "reward": 0.8059885501861572, "reward_std": 0.5008499622344971, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3619750142097473, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 266.0, "completions/max_terminated_length": 266.0, "completions/mean_length": 219.9375, "completions/mean_terminated_length": 219.9375, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.1138, "grad_norm": 0.09663896262645721, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6566740.0, "reward": 1.321191668510437, "reward_std": 0.383238285779953, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8298813104629517, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 873.3125, "completions/mean_terminated_length": 804.8181762695312, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.114, "grad_norm": 0.06340908259153366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6585929.0, "reward": 0.36975377798080444, "reward_std": 0.5527404546737671, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23950624465942383, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.4300000071525574, "step": 570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 795.5, "completions/mean_terminated_length": 762.857177734375, "completions/min_length": 500.0, "completions/min_terminated_length": 500.0, "epoch": 0.1142, "grad_norm": 0.041512615978717804, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6602833.0, "reward": 0.49053525924682617, "reward_std": 0.34253132343292236, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23106873035430908, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6143749952316284, "step": 571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 495.625, "completions/mean_terminated_length": 495.625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.1144, "grad_norm": 0.07066941261291504, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6615019.0, "reward": 0.6178914308547974, "reward_std": 0.4073374271392822, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4857812523841858, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3893750309944153, "step": 572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 405.875, "completions/mean_terminated_length": 405.875, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.1146, "grad_norm": 0.1120079830288887, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6625353.0, "reward": 1.1587947607040405, "reward_std": 0.6214038133621216, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7550874948501587, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 387.9375, "completions/mean_terminated_length": 387.9375, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.1148, "grad_norm": 0.09562289714813232, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6638008.0, "reward": 0.6989852786064148, "reward_std": 0.41589486598968506, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27296876907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 277.25, "completions/mean_terminated_length": 277.25, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.115, "grad_norm": 0.12008745223283768, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6646412.0, "reward": 0.5499169230461121, "reward_std": 0.4263148903846741, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1623312532901764, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 325.125, "completions/mean_terminated_length": 325.125, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1152, "grad_norm": 0.09359157830476761, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6657214.0, "reward": 1.0166103839874268, "reward_std": 0.49844545125961304, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5332187414169312, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 338.9375, "completions/mean_terminated_length": 338.9375, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1154, "grad_norm": 0.07794512808322906, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6666413.0, "reward": 0.6146885752677917, "reward_std": 0.3695681393146515, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2918750047683716, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.737500011920929, "step": 577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 903.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 735.875, "completions/mean_terminated_length": 735.875, "completions/min_length": 498.0, "completions/min_terminated_length": 498.0, "epoch": 0.1156, "grad_norm": 0.04265524074435234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6681963.0, "reward": 0.42507898807525635, "reward_std": 0.25701233744621277, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10015624761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8774999380111694, "step": 578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 376.6875, "completions/mean_terminated_length": 376.6875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.1158, "grad_norm": 0.0025513875298202038, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6692502.0, "reward": 0.5445165634155273, "reward_std": 0.011709420010447502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 529.875, "completions/mean_terminated_length": 459.2857360839844, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.116, "grad_norm": 0.043094467371702194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6708020.0, "reward": 0.4762507975101471, "reward_std": 0.24147053062915802, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14000000059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 687.625, "completions/mean_terminated_length": 665.2000122070312, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.1162, "grad_norm": 0.06150928512215614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6723054.0, "reward": 0.42437615990638733, "reward_std": 0.494486540555954, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16124999523162842, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 454.625, "completions/mean_terminated_length": 416.66668701171875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.1164, "grad_norm": 0.08264277875423431, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6734136.0, "reward": 0.6401569843292236, "reward_std": 0.4434491991996765, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3428124785423279, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.612500011920929, "step": 582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 243.0, "completions/max_terminated_length": 243.0, "completions/mean_length": 180.5625, "completions/mean_terminated_length": 180.5625, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.1166, "grad_norm": 0.1304844170808792, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6740545.0, "reward": 0.7051979899406433, "reward_std": 0.3944871425628662, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22289374470710754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781249761581421, "step": 583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 662.375, "completions/mean_terminated_length": 610.7142944335938, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.1168, "grad_norm": 0.051703087985515594, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6755543.0, "reward": 0.33898240327835083, "reward_std": 0.2759983241558075, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.052962496876716614, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7800000309944153, "step": 584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 419.75, "completions/mean_terminated_length": 379.4666748046875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.117, "grad_norm": 0.1483389139175415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6766659.0, "reward": 0.8501543998718262, "reward_std": 0.6209211349487305, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4503062665462494, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9118750095367432, "step": 585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 789.5625, "completions/mean_terminated_length": 756.0714721679688, "completions/min_length": 560.0, "completions/min_terminated_length": 560.0, "epoch": 0.1172, "grad_norm": 0.06386902183294296, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6783196.0, "reward": 0.6628788709640503, "reward_std": 0.5480340123176575, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3257562518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7768750190734863, "step": 586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 323.4375, "completions/mean_terminated_length": 323.4375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1174, "grad_norm": 0.06998807191848755, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6792307.0, "reward": 1.315685749053955, "reward_std": 0.39509737491607666, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8188687562942505, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756249785423279, "step": 587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 489.0, "completions/max_terminated_length": 489.0, "completions/mean_length": 375.5, "completions/mean_terminated_length": 375.5, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1176, "grad_norm": 0.08335471153259277, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6801995.0, "reward": 1.0214571952819824, "reward_std": 0.49330708384513855, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.542912483215332, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 432.9375, "completions/mean_terminated_length": 432.9375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.1178, "grad_norm": 0.0816422700881958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6813210.0, "reward": 1.0350792407989502, "reward_std": 0.5197245478630066, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6326562166213989, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 993.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 579.8125, "completions/mean_terminated_length": 579.8125, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.118, "grad_norm": 0.056763097643852234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6828279.0, "reward": 0.48390722274780273, "reward_std": 0.32679393887519836, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21781250834465027, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 355.0625, "completions/mean_terminated_length": 355.0625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.1182, "grad_norm": 0.10358447581529617, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6838360.0, "reward": 1.1674103736877441, "reward_std": 0.4549490809440613, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7098187208175659, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 331.0625, "completions/mean_terminated_length": 331.0625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.1184, "grad_norm": 0.09448617696762085, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6847513.0, "reward": 0.8312947750091553, "reward_std": 0.46553856134414673, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3500875234603882, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 711.9375, "completions/mean_terminated_length": 711.9375, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.1186, "grad_norm": 0.06674634665250778, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6863704.0, "reward": 0.5903131365776062, "reward_std": 0.43821439146995544, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49312499165534973, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.296875, "step": 593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 387.5, "completions/mean_terminated_length": 387.5, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.1188, "grad_norm": 0.09204850345849991, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6873904.0, "reward": 1.0203102827072144, "reward_std": 0.4591692090034485, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6031187772750854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8493750095367432, "step": 594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 401.375, "completions/mean_terminated_length": 401.375, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.119, "grad_norm": 0.07454129308462143, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6884630.0, "reward": 0.8409010171890259, "reward_std": 0.4585542380809784, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3693000078201294, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 547.4375, "completions/mean_terminated_length": 515.6666870117188, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1192, "grad_norm": 0.06179853528738022, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6898205.0, "reward": 0.7852290272712708, "reward_std": 0.4309122860431671, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.44545626640319824, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7268750071525574, "step": 596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 523.125, "completions/mean_terminated_length": 523.125, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.1194, "grad_norm": 0.07579565793275833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6911135.0, "reward": 0.7535165548324585, "reward_std": 0.49932247400283813, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38203126192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 487.0625, "completions/mean_terminated_length": 487.0625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.1196, "grad_norm": 0.04614569991827011, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6922288.0, "reward": 0.7527349591255188, "reward_std": 0.22541998326778412, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5679687261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.609375, "step": 598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 397.0, "completions/mean_terminated_length": 397.0, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.1198, "grad_norm": 0.052408795803785324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6932896.0, "reward": 1.380751132965088, "reward_std": 0.32478007674217224, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8865000009536743, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 268.875, "completions/mean_terminated_length": 268.875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.12, "grad_norm": 0.06754652410745621, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6941230.0, "reward": 0.59747314453125, "reward_std": 0.2461005449295044, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13244375586509705, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 492.0, "completions/mean_terminated_length": 492.0, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.1202, "grad_norm": 0.09084391593933105, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6953166.0, "reward": 0.8603604435920715, "reward_std": 0.5582209825515747, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47071874141693115, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 569.0, "completions/mean_terminated_length": 569.0, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1204, "grad_norm": 0.006973578128963709, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6967550.0, "reward": 0.5556010007858276, "reward_std": 0.04784756898880005, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1111999899148941, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9412499666213989, "step": 602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 710.4375, "completions/mean_terminated_length": 638.0769653320312, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.1206, "grad_norm": 0.062409307807683945, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6982789.0, "reward": 0.49921953678131104, "reward_std": 0.5016158819198608, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24843749403953552, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6156249642372131, "step": 603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 373.5, "completions/mean_terminated_length": 373.5, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.1208, "grad_norm": 0.10511057078838348, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6993117.0, "reward": 1.3147542476654053, "reward_std": 0.43633556365966797, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8795062303543091, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 409.375, "completions/mean_terminated_length": 409.375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.121, "grad_norm": 0.0397895947098732, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7004163.0, "reward": 0.47653841972351074, "reward_std": 0.18627846240997314, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07807499170303345, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 652.5625, "completions/mean_terminated_length": 627.800048828125, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.1212, "grad_norm": 0.0652690902352333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7018684.0, "reward": 0.7102883458137512, "reward_std": 0.528262734413147, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35807499289512634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.862500011920929, "step": 606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 422.3125, "completions/mean_terminated_length": 422.3125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.1214, "grad_norm": 0.00413115369156003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7029873.0, "reward": 0.5305637121200562, "reward_std": 0.020045671612024307, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.061124999076128006, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 830.75, "completions/mean_terminated_length": 766.3333740234375, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.1216, "grad_norm": 0.05198018252849579, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7047325.0, "reward": 0.4278883934020996, "reward_std": 0.4499684274196625, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23077498376369476, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.6106250286102295, "step": 608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 414.4375, "completions/mean_terminated_length": 414.4375, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.1218, "grad_norm": 0.08921852707862854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7057780.0, "reward": 0.8693759441375732, "reward_std": 0.48720699548721313, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.48874998092651367, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "step": 609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 452.0, "completions/max_terminated_length": 452.0, "completions/mean_length": 387.0625, "completions/mean_terminated_length": 387.0625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.122, "grad_norm": 0.08377812057733536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7068021.0, "reward": 1.316535472869873, "reward_std": 0.3938465118408203, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8205687999725342, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 352.25, "completions/mean_terminated_length": 352.25, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.1222, "grad_norm": 0.08888816833496094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7077785.0, "reward": 0.6691854000091553, "reward_std": 0.4315739870071411, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21336837112903595, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9797500371932983, "step": 611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 428.25, "completions/mean_terminated_length": 428.25, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1224, "grad_norm": 0.060103558003902435, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7089181.0, "reward": 1.026238203048706, "reward_std": 0.328782320022583, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8024749755859375, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5237500071525574, "step": 612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 350.1875, "completions/mean_terminated_length": 350.1875, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.1226, "grad_norm": 0.07467937469482422, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7098656.0, "reward": 1.346001148223877, "reward_std": 0.3440401554107666, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8795000314712524, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9075000286102295, "step": 613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 354.0625, "completions/mean_terminated_length": 354.0625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1228, "grad_norm": 0.05734195560216904, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7108433.0, "reward": 0.6763821840286255, "reward_std": 0.3241499960422516, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22776249051094055, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9412499666213989, "step": 614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 252.125, "completions/mean_terminated_length": 252.125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.123, "grad_norm": 0.04264270141720772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7116483.0, "reward": 0.5011353492736816, "reward_std": 0.13479578495025635, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06476874649524689, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 305.5625, "completions/mean_terminated_length": 305.5625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1232, "grad_norm": 0.11375895142555237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7125116.0, "reward": 1.1631667613983154, "reward_std": 0.5294799208641052, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7013312578201294, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 622.75, "completions/mean_terminated_length": 565.4285888671875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1234, "grad_norm": 0.05281246453523636, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7139032.0, "reward": 0.49451982975006104, "reward_std": 0.37742137908935547, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17653748393058777, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8112500309944153, "step": 617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 512.625, "completions/mean_terminated_length": 512.625, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.1236, "grad_norm": 0.12584543228149414, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7152434.0, "reward": 0.8837919235229492, "reward_std": 0.6652922630310059, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5175812244415283, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.971250057220459, "step": 618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 536.0625, "completions/mean_terminated_length": 423.4615478515625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.1238, "grad_norm": 0.10034535825252533, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7165571.0, "reward": 0.8228100538253784, "reward_std": 0.6222297549247742, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.520618736743927, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6775000095367432, "step": 619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 451.625, "completions/mean_terminated_length": 451.625, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.124, "grad_norm": 0.08021686971187592, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7177117.0, "reward": 0.7816355228424072, "reward_std": 0.42835983633995056, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3132687509059906, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 350.6875, "completions/mean_terminated_length": 350.6875, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1242, "grad_norm": 0.16049490869045258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7187560.0, "reward": 1.1537511348724365, "reward_std": 0.5646336078643799, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7450000047683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9281249642372131, "step": 621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 428.0, "completions/mean_terminated_length": 428.0, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.1244, "grad_norm": 0.1006806343793869, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7198072.0, "reward": 1.0166759490966797, "reward_std": 0.49851322174072266, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5333499908447266, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 185.1875, "completions/mean_terminated_length": 185.1875, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.1246, "grad_norm": 0.148814395070076, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7204907.0, "reward": 0.7019981741905212, "reward_std": 0.39429914951324463, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21649375557899475, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 714.5, "completions/mean_terminated_length": 670.2857666015625, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.1248, "grad_norm": 0.057843055576086044, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7221027.0, "reward": 0.8025789856910706, "reward_std": 0.5713921189308167, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4176562428474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7593749761581421, "step": 624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 607.75, "completions/mean_terminated_length": 548.2857666015625, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.125, "grad_norm": 0.04599748179316521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7235295.0, "reward": 0.5408599376678467, "reward_std": 0.3682137429714203, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3317187428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.581250011920929, "step": 625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 410.375, "completions/mean_terminated_length": 410.375, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1252, "grad_norm": 0.06244147941470146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7245861.0, "reward": 0.6530351638793945, "reward_std": 0.30827397108078003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3060687482357025, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7831250429153442, "step": 626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 235.0, "completions/max_terminated_length": 235.0, "completions/mean_length": 194.375, "completions/mean_terminated_length": 194.375, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.1254, "grad_norm": 0.15148134529590607, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7252715.0, "reward": 1.2767698764801025, "reward_std": 0.40648016333580017, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8035374879837036, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 594.4375, "completions/mean_terminated_length": 565.800048828125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.1256, "grad_norm": 0.10664805769920349, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7266770.0, "reward": 0.7917193174362183, "reward_std": 0.5357744097709656, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6459375023841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.484375, "step": 628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 592.5625, "completions/mean_terminated_length": 592.5625, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.1258, "grad_norm": 0.04393412545323372, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7280331.0, "reward": 0.5703134536743164, "reward_std": 0.26839691400527954, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.265625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 389.0625, "completions/mean_terminated_length": 346.73333740234375, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.126, "grad_norm": 0.08692184090614319, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7290620.0, "reward": 0.7749883532524109, "reward_std": 0.4686538577079773, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3624750077724457, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 412.5625, "completions/mean_terminated_length": 412.5625, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.1262, "grad_norm": 0.017604952678084373, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7301173.0, "reward": 0.5883479118347168, "reward_std": 0.08390381932258606, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17669375240802765, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 564.4375, "completions/mean_terminated_length": 564.4375, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.1264, "grad_norm": 0.05950119346380234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7314332.0, "reward": 1.2544729709625244, "reward_std": 0.4382792115211487, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7589437365531921, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 357.75, "completions/mean_terminated_length": 357.75, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1266, "grad_norm": 0.09109313040971756, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7323768.0, "reward": 0.6852351427078247, "reward_std": 0.4152469038963318, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.37046873569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 364.6875, "completions/mean_terminated_length": 364.6875, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.1268, "grad_norm": 0.08529230952262878, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7333827.0, "reward": 0.7941417098045349, "reward_std": 0.4573298990726471, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.2757812440395355, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.778124988079071, "step": 634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 606.9375, "completions/mean_terminated_length": 547.357177734375, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.127, "grad_norm": 0.02437669038772583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7347282.0, "reward": 0.4606948494911194, "reward_std": 0.18114735186100006, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04638749733567238, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8512500524520874, "step": 635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 780.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 416.5, "completions/mean_terminated_length": 416.5, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1272, "grad_norm": 0.04598161578178406, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7357706.0, "reward": 0.5909324884414673, "reward_std": 0.2437298446893692, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11936250329017639, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 856.625, "completions/mean_terminated_length": 800.8333740234375, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.1274, "grad_norm": 0.04237482324242592, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7376084.0, "reward": 0.3007037937641144, "reward_std": 0.36906155943870544, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1639062464237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.559374988079071, "step": 637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 344.25, "completions/mean_terminated_length": 344.25, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.1276, "grad_norm": 0.09143860638141632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7385528.0, "reward": 1.0178918838500977, "reward_std": 0.4973527491092682, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5357812643051147, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 700.3125, "completions/mean_terminated_length": 654.0714721679688, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.1278, "grad_norm": 0.0381949283182621, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7400621.0, "reward": 0.6231257915496826, "reward_std": 0.35311535000801086, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3087500035762787, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 322.4375, "completions/mean_terminated_length": 322.4375, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.128, "grad_norm": 0.12122191488742828, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7409380.0, "reward": 1.1082696914672852, "reward_std": 0.5366030335426331, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6540374755859375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 380.4375, "completions/mean_terminated_length": 380.4375, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.1282, "grad_norm": 0.08591371774673462, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7419147.0, "reward": 1.086225986480713, "reward_std": 0.48366233706474304, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.609950065612793, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 597.4375, "completions/mean_terminated_length": 569.0, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1284, "grad_norm": 0.07068858295679092, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7432594.0, "reward": 0.8654509782791138, "reward_std": 0.5242062211036682, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41840001940727234, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500619888306, "step": 642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 302.3125, "completions/mean_terminated_length": 302.3125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1286, "grad_norm": 0.11376109719276428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7441591.0, "reward": 0.8091009259223938, "reward_std": 0.4991062879562378, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36820000410079956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 348.875, "completions/mean_terminated_length": 348.875, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1288, "grad_norm": 0.09580470621585846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7452213.0, "reward": 1.2599196434020996, "reward_std": 0.4284239709377289, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.769837498664856, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 370.1875, "completions/mean_terminated_length": 370.1875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.129, "grad_norm": 0.16799162328243256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7461848.0, "reward": 0.8822231292724609, "reward_std": 0.5867282748222351, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4519437551498413, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 900.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 599.875, "completions/mean_terminated_length": 599.875, "completions/min_length": 464.0, "completions/min_terminated_length": 464.0, "epoch": 0.1292, "grad_norm": 0.050109222531318665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7476294.0, "reward": 0.5581074953079224, "reward_std": 0.28467321395874023, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11621250212192535, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 552.0625, "completions/mean_terminated_length": 552.0625, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.1294, "grad_norm": 0.05674540251493454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7488807.0, "reward": 0.6173696517944336, "reward_std": 0.4035293459892273, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23473750054836273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9206249713897705, "step": 647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 422.25, "completions/mean_terminated_length": 422.25, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.1296, "grad_norm": 0.04172791913151741, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7499867.0, "reward": 0.6032697558403015, "reward_std": 0.23918719589710236, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14403750002384186, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 434.625, "completions/mean_terminated_length": 434.625, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.1298, "grad_norm": 0.06578578054904938, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7510309.0, "reward": 1.380663514137268, "reward_std": 0.32474592328071594, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8863250017166138, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 400.875, "completions/mean_terminated_length": 400.875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.13, "grad_norm": 0.0813649520277977, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7520899.0, "reward": 0.9646071195602417, "reward_std": 0.4862818419933319, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4917125105857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 787.4375, "completions/mean_terminated_length": 771.6666870117188, "completions/min_length": 560.0, "completions/min_terminated_length": 560.0, "epoch": 0.1302, "grad_norm": 0.06825343519449234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7537690.0, "reward": 0.5236571431159973, "reward_std": 0.48478031158447266, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23481249809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6918750405311584, "step": 651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 704.8125, "completions/mean_terminated_length": 683.5333862304688, "completions/min_length": 532.0, "completions/min_terminated_length": 532.0, "epoch": 0.1304, "grad_norm": 0.10387086868286133, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7552903.0, "reward": 0.8501384258270264, "reward_std": 0.6471896767616272, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5127747654914856, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8461250066757202, "step": 652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 259.125, "completions/mean_terminated_length": 259.125, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.1306, "grad_norm": 0.09730765223503113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7561753.0, "reward": 0.7699136734008789, "reward_std": 0.4352329671382904, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2898250222206116, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 373.6875, "completions/mean_terminated_length": 373.6875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.1308, "grad_norm": 0.07231306284666061, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7571780.0, "reward": 0.7696100473403931, "reward_std": 0.328118234872818, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6642187833786011, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.41874998807907104, "step": 654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 287.1875, "completions/mean_terminated_length": 287.1875, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.131, "grad_norm": 0.091468945145607, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7580135.0, "reward": 0.655288577079773, "reward_std": 0.3304465413093567, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18557500839233398, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 266.0, "completions/max_terminated_length": 266.0, "completions/mean_length": 180.625, "completions/mean_terminated_length": 180.625, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.1312, "grad_norm": 0.0418376587331295, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7586993.0, "reward": 1.4370856285095215, "reward_std": 0.17062129080295563, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9991687536239624, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 700.0, "completions/mean_terminated_length": 700.0, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.1314, "grad_norm": 0.037334226071834564, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7602785.0, "reward": 0.572091817855835, "reward_std": 0.2330905944108963, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2066812366247177, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8287500143051147, "step": 657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 469.0625, "completions/mean_terminated_length": 469.0625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.1316, "grad_norm": 0.007674560882151127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7615554.0, "reward": 0.5584385991096497, "reward_std": 0.056526895612478256, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11687499284744263, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 611.0625, "completions/mean_terminated_length": 611.0625, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.1318, "grad_norm": 0.0738237276673317, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7629395.0, "reward": 0.8293574452400208, "reward_std": 0.5638657808303833, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4087125062942505, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9100000262260437, "step": 659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 640.0625, "completions/mean_terminated_length": 614.4666748046875, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.132, "grad_norm": 0.022772708907723427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7644020.0, "reward": 0.5034384727478027, "reward_std": 0.21259966492652893, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13187499344348907, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 558.5625, "completions/mean_terminated_length": 558.5625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.1322, "grad_norm": 0.07922293245792389, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7656861.0, "reward": 0.9672665596008301, "reward_std": 0.4842281639575958, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49703124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 461.4375, "completions/mean_terminated_length": 461.4375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.1324, "grad_norm": 0.07007812708616257, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7668484.0, "reward": 0.7269415855407715, "reward_std": 0.3830983638763428, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26638123393058777, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 339.0, "completions/max_terminated_length": 339.0, "completions/mean_length": 285.125, "completions/mean_terminated_length": 285.125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1326, "grad_norm": 0.15570120513439178, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7676982.0, "reward": 1.1492384672164917, "reward_std": 0.4688589870929718, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6734750270843506, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 436.375, "completions/mean_terminated_length": 436.375, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1328, "grad_norm": 0.07444309443235397, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7687740.0, "reward": 0.9276571273803711, "reward_std": 0.4177421033382416, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5428124666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 362.125, "completions/mean_terminated_length": 362.125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.133, "grad_norm": 0.14346350729465485, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7697934.0, "reward": 0.7992199659347534, "reward_std": 0.5956372618675232, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41093748807907104, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 501.8125, "completions/mean_terminated_length": 501.8125, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.1332, "grad_norm": 0.09568187594413757, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7710427.0, "reward": 0.8883887529373169, "reward_std": 0.5828195214271545, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4642750024795532, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 337.125, "completions/mean_terminated_length": 337.125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1334, "grad_norm": 0.051451053470373154, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7719677.0, "reward": 0.6903914213180542, "reward_std": 0.2832822799682617, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.38078123331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7937500476837158, "step": 667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 531.3125, "completions/mean_terminated_length": 531.3125, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.1336, "grad_norm": 0.055549152195453644, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7732514.0, "reward": 1.3191447257995605, "reward_std": 0.3870584964752197, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8257875442504883, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 427.0, "completions/mean_terminated_length": 427.0, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1338, "grad_norm": 0.05951762944459915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7743346.0, "reward": 0.5103079080581665, "reward_std": 0.31655725836753845, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08311250060796738, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9862500429153442, "step": 669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 450.3125, "completions/mean_terminated_length": 450.3125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.134, "grad_norm": 0.10547398775815964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7755559.0, "reward": 1.2274949550628662, "reward_std": 0.5017624497413635, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7674875259399414, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 345.875, "completions/mean_terminated_length": 345.875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.1342, "grad_norm": 0.059593915939331055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7765477.0, "reward": 0.6597664952278137, "reward_std": 0.24715003371238708, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.25703126192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 305.0, "completions/max_terminated_length": 305.0, "completions/mean_length": 222.625, "completions/mean_terminated_length": 222.625, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.1344, "grad_norm": 0.10939165204763412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7772703.0, "reward": 1.320997953414917, "reward_std": 0.3831422030925751, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8294937610626221, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.1346, "grad_norm": 0.08248494565486908, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7782497.0, "reward": 1.2595821619033813, "reward_std": 0.42822256684303284, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691624760627747, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 295.6875, "completions/mean_terminated_length": 295.6875, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.1348, "grad_norm": 0.08747648447751999, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7791468.0, "reward": 1.3208353519439697, "reward_std": 0.38306140899658203, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291687965393066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 535.4375, "completions/mean_terminated_length": 535.4375, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.135, "grad_norm": 0.06656351685523987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7804019.0, "reward": 0.9914853572845459, "reward_std": 0.43195903301239014, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6079687476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 517.625, "completions/mean_terminated_length": 517.625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1352, "grad_norm": 0.04696691781282425, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7816125.0, "reward": 0.6532821655273438, "reward_std": 0.22031302750110626, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24406248331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 540.6875, "completions/mean_terminated_length": 540.6875, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.1354, "grad_norm": 0.07479993999004364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7828552.0, "reward": 0.8305481672286987, "reward_std": 0.5005719661712646, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4110937714576721, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.846875011920929, "step": 677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 425.5625, "completions/mean_terminated_length": 425.5625, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.1356, "grad_norm": 0.08511261641979218, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7839857.0, "reward": 0.7717168927192688, "reward_std": 0.43420180678367615, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29343125224113464, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 276.0, "completions/max_terminated_length": 276.0, "completions/mean_length": 232.125, "completions/mean_terminated_length": 232.125, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.1358, "grad_norm": 0.08660000562667847, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7847395.0, "reward": 1.4397573471069336, "reward_std": 0.23760239779949188, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9420124888420105, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 503.1875, "completions/mean_terminated_length": 503.1875, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.136, "grad_norm": 0.08560966700315475, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7859062.0, "reward": 1.0780010223388672, "reward_std": 0.4936424493789673, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.593500018119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 359.0625, "completions/mean_terminated_length": 359.0625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1362, "grad_norm": 0.09306765347719193, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7868375.0, "reward": 1.0142604112625122, "reward_std": 0.5010389089584351, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.528518795967102, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 279.0625, "completions/mean_terminated_length": 279.0625, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.1364, "grad_norm": 0.05132042244076729, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7876424.0, "reward": 1.4395822286605835, "reward_std": 0.2375580221414566, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416625499725342, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 201.0, "completions/max_terminated_length": 201.0, "completions/mean_length": 169.25, "completions/mean_terminated_length": 169.25, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.1366, "grad_norm": 0.17485623061656952, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7882796.0, "reward": 1.1352012157440186, "reward_std": 0.48619380593299866, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6454000473022461, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 973.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 545.625, "completions/mean_terminated_length": 545.625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1368, "grad_norm": 0.0867292657494545, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7895190.0, "reward": 1.0710946321487427, "reward_std": 0.45803406834602356, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6421874761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 331.0625, "completions/mean_terminated_length": 331.0625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.137, "grad_norm": 0.1660027652978897, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7904135.0, "reward": 0.7737510204315186, "reward_std": 0.5354137420654297, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36000001430511475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.949999988079071, "step": 685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 538.4375, "completions/mean_terminated_length": 506.0666809082031, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.1372, "grad_norm": 0.08068744093179703, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7916734.0, "reward": 0.7260946035385132, "reward_std": 0.36584270000457764, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.20218749344348907, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.625, "step": 686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 557.4375, "completions/mean_terminated_length": 526.3333740234375, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.1374, "grad_norm": 0.0916156992316246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7929909.0, "reward": 0.8960853815078735, "reward_std": 0.5772761702537537, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4796687662601471, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9043750762939453, "step": 687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 261.875, "completions/mean_terminated_length": 261.875, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.1376, "grad_norm": 0.09211242198944092, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7938755.0, "reward": 1.3187323808670044, "reward_std": 0.3885209858417511, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8249624967575073, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 399.0625, "completions/mean_terminated_length": 399.0625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.1378, "grad_norm": 0.10579158365726471, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7949172.0, "reward": 0.7080978155136108, "reward_std": 0.5043999552726746, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29119372367858887, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9075000286102295, "step": 689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 414.0, "completions/mean_terminated_length": 414.0, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.138, "grad_norm": 0.0020358695182949305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7960116.0, "reward": 0.551641583442688, "reward_std": 0.011562498286366463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10328124463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 600.875, "completions/mean_terminated_length": 572.6666870117188, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.1382, "grad_norm": 0.06959745287895203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7974802.0, "reward": 0.6865636110305786, "reward_std": 0.4251965284347534, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24812498688697815, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 526.75, "completions/mean_terminated_length": 493.60003662109375, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1384, "grad_norm": 0.09080769121646881, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7987262.0, "reward": 0.9218606948852539, "reward_std": 0.5425811409950256, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4687187671661377, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9168750047683716, "step": 692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 805.6875, "completions/mean_terminated_length": 674.7000122070312, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.1386, "grad_norm": 0.06890544295310974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8005417.0, "reward": 0.4650786221027374, "reward_std": 0.5239605903625488, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30515626072883606, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.3062500059604645, "step": 693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 414.375, "completions/mean_terminated_length": 414.375, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.1388, "grad_norm": 0.13535702228546143, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8016383.0, "reward": 1.1655009984970093, "reward_std": 0.5262821912765503, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7059999704360962, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 704.1875, "completions/mean_terminated_length": 630.3846435546875, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.139, "grad_norm": 0.07327032089233398, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8031954.0, "reward": 0.5268758535385132, "reward_std": 0.5483844876289368, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24125000834465027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 433.8125, "completions/mean_terminated_length": 433.8125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1392, "grad_norm": 0.058233246207237244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8042863.0, "reward": 0.5493760108947754, "reward_std": 0.32141998410224915, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16124999523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9375, "step": 696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 430.625, "completions/mean_terminated_length": 430.625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.1394, "grad_norm": 0.07754460722208023, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8054521.0, "reward": 0.577735424041748, "reward_std": 0.34552907943725586, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21796874701976776, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.765625, "step": 697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 718.8125, "completions/mean_terminated_length": 718.8125, "completions/min_length": 547.0, "completions/min_terminated_length": 547.0, "epoch": 0.1396, "grad_norm": 0.06744838505983353, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8070342.0, "reward": 0.584913432598114, "reward_std": 0.5096649527549744, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23232501745224, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 290.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 222.5625, "completions/mean_terminated_length": 222.5625, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.1398, "grad_norm": 0.0069003296084702015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8077439.0, "reward": 0.5305822491645813, "reward_std": 0.019021349027752876, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06116250157356262, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 741.5, "completions/mean_terminated_length": 701.1428833007812, "completions/min_length": 523.0, "completions/min_terminated_length": 523.0, "epoch": 0.14, "grad_norm": 0.06446799635887146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8093591.0, "reward": 0.4914068579673767, "reward_std": 0.3870856463909149, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2953125238418579, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.4437499940395355, "step": 700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 357.6875, "completions/mean_terminated_length": 357.6875, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1402, "grad_norm": 0.04689962416887283, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8103218.0, "reward": 0.5764853954315186, "reward_std": 0.24556121230125427, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21546855568885803, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8042812347412109, "step": 701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 346.0, "completions/mean_terminated_length": 346.0, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.1404, "grad_norm": 0.09404025226831436, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8112978.0, "reward": 0.8900008201599121, "reward_std": 0.35244274139404297, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5924999713897705, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6624999642372131, "step": 702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 471.0, "completions/max_terminated_length": 471.0, "completions/mean_length": 292.1875, "completions/mean_terminated_length": 292.1875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.1406, "grad_norm": 0.15146689116954803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8121093.0, "reward": 1.076063632965088, "reward_std": 0.4956435561180115, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5896250009536743, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 636.875, "completions/mean_terminated_length": 547.5384521484375, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.1408, "grad_norm": 0.08216100931167603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8134835.0, "reward": 0.7278102040290833, "reward_std": 0.600522518157959, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3931187689304352, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7293750047683716, "step": 704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 908.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 550.625, "completions/mean_terminated_length": 550.625, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.141, "grad_norm": 0.09373030066490173, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8147261.0, "reward": 0.8845919370651245, "reward_std": 0.5851024389266968, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4566812515258789, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9181250333786011, "step": 705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 280.75, "completions/mean_terminated_length": 280.75, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1412, "grad_norm": 0.00015780395187903196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8156185.0, "reward": 1.4994230270385742, "reward_std": 0.0006136074080131948, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998843789100647, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 324.0, "completions/max_terminated_length": 324.0, "completions/mean_length": 264.0625, "completions/mean_terminated_length": 264.0625, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.1414, "grad_norm": 0.06460901349782944, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8164122.0, "reward": 0.759644627571106, "reward_std": 0.23765958845615387, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5192874670028687, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5362499952316284, "step": 707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 333.4375, "completions/mean_terminated_length": 333.4375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.1416, "grad_norm": 0.003224818967282772, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8173345.0, "reward": 0.5545321702957153, "reward_std": 0.015797382220625877, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10906249284744263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 363.6875, "completions/mean_terminated_length": 363.6875, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.1418, "grad_norm": 0.11577444523572922, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8183116.0, "reward": 1.1905417442321777, "reward_std": 0.5687910318374634, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.756081223487854, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 370.5, "completions/mean_terminated_length": 326.933349609375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.142, "grad_norm": 0.1140500009059906, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8192948.0, "reward": 0.9871071577072144, "reward_std": 0.5447024703025818, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5367125272750854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 888.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 437.4375, "completions/mean_terminated_length": 437.4375, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1422, "grad_norm": 0.06752481311559677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8204427.0, "reward": 0.7437353134155273, "reward_std": 0.32185184955596924, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3624686896800995, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8343124985694885, "step": 711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 420.0625, "completions/mean_terminated_length": 420.0625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.1424, "grad_norm": 0.07930604368448257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8215548.0, "reward": 0.9033353328704834, "reward_std": 0.4766920804977417, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.431668758392334, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 509.0625, "completions/mean_terminated_length": 509.0625, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.1426, "grad_norm": 0.0893934965133667, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8227725.0, "reward": 0.890191912651062, "reward_std": 0.4879489243030548, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4053812325000763, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 471.0625, "completions/mean_terminated_length": 471.0625, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.1428, "grad_norm": 0.07721509039402008, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8238974.0, "reward": 0.6163290739059448, "reward_std": 0.4404827654361725, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.29515624046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 702.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 421.3125, "completions/mean_terminated_length": 421.3125, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.143, "grad_norm": 0.08180579543113708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8249299.0, "reward": 0.9292949438095093, "reward_std": 0.45786744356155396, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.48358750343322754, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8650000095367432, "step": 715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 646.9375, "completions/mean_terminated_length": 621.800048828125, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.1432, "grad_norm": 0.07596620172262192, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8264034.0, "reward": 1.1585793495178223, "reward_std": 0.6212787628173828, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.754656195640564, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8481250405311584, "step": 716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 805.6875, "completions/mean_terminated_length": 674.7000122070312, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.1434, "grad_norm": 0.041289281100034714, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8282125.0, "reward": 0.3407692313194275, "reward_std": 0.321880578994751, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11903749406337738, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4987500011920929, "step": 717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 463.625, "completions/mean_terminated_length": 463.625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1436, "grad_norm": 0.0720892921090126, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8293639.0, "reward": 0.8959386944770813, "reward_std": 0.4827613830566406, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4168750047683716, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 508.5, "completions/mean_terminated_length": 508.5, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.1438, "grad_norm": 0.014072166755795479, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8306047.0, "reward": 0.5311734676361084, "reward_std": 0.06395834684371948, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.062343746423721313, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 880.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 447.9375, "completions/mean_terminated_length": 447.9375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.144, "grad_norm": 0.12331701070070267, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8317438.0, "reward": 0.9146450161933899, "reward_std": 0.5741459727287292, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5167874693870544, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9300000071525574, "step": 720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 263.5625, "completions/mean_terminated_length": 263.5625, "completions/min_length": 137.0, "completions/min_terminated_length": 137.0, "epoch": 0.1442, "grad_norm": 0.11253634840250015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8325463.0, "reward": 1.1265606880187988, "reward_std": 0.4816039204597473, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6281187534332275, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 833.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 555.75, "completions/mean_terminated_length": 555.75, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.1444, "grad_norm": 0.07465104013681412, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8338291.0, "reward": 0.9563041925430298, "reward_std": 0.45533958077430725, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5376062393188477, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8756250143051147, "step": 722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 294.4375, "completions/mean_terminated_length": 294.4375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.1446, "grad_norm": 0.08805017918348312, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8347546.0, "reward": 1.346072793006897, "reward_std": 0.4301775395870209, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8796437978744507, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9343750476837158, "step": 723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 633.8125, "completions/mean_terminated_length": 607.800048828125, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.1448, "grad_norm": 0.03301485255360603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8361527.0, "reward": 0.7231255769729614, "reward_std": 0.24735260009765625, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5087500214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5874999761581421, "step": 724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 451.0625, "completions/mean_terminated_length": 451.0625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.145, "grad_norm": 0.10640236735343933, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8372664.0, "reward": 0.5602350234985352, "reward_std": 0.45831558108329773, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49546876549720764, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5249999761581421, "step": 725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 777.6875, "completions/mean_terminated_length": 742.5000610351562, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.1452, "grad_norm": 0.06284786015748978, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8390275.0, "reward": 0.588047981262207, "reward_std": 0.6018084287643433, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3010937571525574, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 380.1875, "completions/mean_terminated_length": 380.1875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.1454, "grad_norm": 0.09564993530511856, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8400486.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 396.375, "completions/mean_terminated_length": 396.375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.1456, "grad_norm": 0.10796830803155899, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8411292.0, "reward": 0.9028134346008301, "reward_std": 0.595345139503479, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5556249618530273, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.706250011920929, "step": 728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 478.25, "completions/mean_terminated_length": 478.25, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.1458, "grad_norm": 0.07700895518064499, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8423952.0, "reward": 0.7537510395050049, "reward_std": 0.4293783903121948, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3824999928474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8362499475479126, "step": 729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 357.875, "completions/mean_terminated_length": 357.875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.146, "grad_norm": 0.10410855710506439, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8433646.0, "reward": 1.2020291090011597, "reward_std": 0.45488473773002625, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7165562510490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 196.0, "completions/max_terminated_length": 196.0, "completions/mean_length": 144.375, "completions/mean_terminated_length": 144.375, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 0.1462, "grad_norm": 0.003417975502088666, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8440164.0, "reward": 0.5031110644340515, "reward_std": 0.004762712400406599, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006218750029802322, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.996874988079071, "step": 731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1015.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 448.1875, "completions/mean_terminated_length": 448.1875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.1464, "grad_norm": 0.0733409970998764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8451751.0, "reward": 0.6602979302406311, "reward_std": 0.3279242217540741, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19559374451637268, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 821.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 573.8125, "completions/mean_terminated_length": 573.8125, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.1466, "grad_norm": 0.04693417623639107, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8464484.0, "reward": 0.6702353954315186, "reward_std": 0.3240903913974762, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21546873450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 385.25, "completions/mean_terminated_length": 385.25, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.1468, "grad_norm": 0.00012374889047350734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8474648.0, "reward": 1.4994478225708008, "reward_std": 0.0005614577676169574, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9988937377929688, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9718750715255737, "step": 734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 381.375, "completions/mean_terminated_length": 381.375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.147, "grad_norm": 0.1034405380487442, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8485518.0, "reward": 0.975469708442688, "reward_std": 0.4777337610721588, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5134375095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 331.5, "completions/mean_terminated_length": 331.5, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.1472, "grad_norm": 0.08606355637311935, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8495126.0, "reward": 0.6377107501029968, "reward_std": 0.33660033345222473, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1504187434911728, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 684.5625, "completions/mean_terminated_length": 661.933349609375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.1474, "grad_norm": 0.061839863657951355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8510479.0, "reward": 0.6334382891654968, "reward_std": 0.37481003999710083, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4543750286102295, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5406249761581421, "step": 737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 325.4375, "completions/mean_terminated_length": 325.4375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.1476, "grad_norm": 0.13952389359474182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8520214.0, "reward": 0.9800355434417725, "reward_std": 0.48719531297683716, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5225687623023987, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 278.0, "completions/mean_terminated_length": 278.0, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.1478, "grad_norm": 0.16208875179290771, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8528918.0, "reward": 1.0239696502685547, "reward_std": 0.4908040165901184, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5479375123977661, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 427.5625, "completions/mean_terminated_length": 427.5625, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.148, "grad_norm": 0.003496797289699316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8539375.0, "reward": 0.5545321702957153, "reward_std": 0.015797382220625877, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10906249284744263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 556.0, "completions/mean_terminated_length": 556.0, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.1482, "grad_norm": 0.06587132811546326, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8553039.0, "reward": 0.6775542497634888, "reward_std": 0.4287927746772766, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23010623455047607, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.971250057220459, "step": 741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 403.4375, "completions/mean_terminated_length": 403.4375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1484, "grad_norm": 0.07555490732192993, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8563622.0, "reward": 0.7585945725440979, "reward_std": 0.41402968764305115, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3921874761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 327.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 223.8125, "completions/mean_terminated_length": 223.8125, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.1486, "grad_norm": 0.16626466810703278, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8571107.0, "reward": 0.8371073603630066, "reward_std": 0.4610399603843689, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3617125153541565, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 694.125, "completions/mean_terminated_length": 647.0, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.1488, "grad_norm": 0.09481579065322876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8586085.0, "reward": 0.7410165071487427, "reward_std": 0.5955180525779724, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4195312559604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.78125, "step": 744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 678.6875, "completions/mean_terminated_length": 655.6666870117188, "completions/min_length": 484.0, "completions/min_terminated_length": 484.0, "epoch": 0.149, "grad_norm": 0.07749906927347183, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8600976.0, "reward": 0.8876665830612183, "reward_std": 0.5829669237136841, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4628312587738037, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9137500524520874, "step": 745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 269.25, "completions/mean_terminated_length": 269.25, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.1492, "grad_norm": 0.14518947899341583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8608884.0, "reward": 1.081763505935669, "reward_std": 0.48930537700653076, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6010249853134155, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 407.6875, "completions/mean_terminated_length": 407.6875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.1494, "grad_norm": 0.10001582652330399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8618879.0, "reward": 0.8660728931427002, "reward_std": 0.5237493515014648, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.419643759727478, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.965624988079071, "step": 747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 482.75, "completions/mean_terminated_length": 446.66668701171875, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.1496, "grad_norm": 0.09582788497209549, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8630571.0, "reward": 1.0474320650100708, "reward_std": 0.5451657772064209, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5948625206947327, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9050000309944153, "step": 748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 730.8125, "completions/mean_terminated_length": 730.8125, "completions/min_length": 633.0, "completions/min_terminated_length": 633.0, "epoch": 0.1498, "grad_norm": 0.025520434603095055, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8646952.0, "reward": 0.5569072961807251, "reward_std": 0.28466394543647766, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11381249129772186, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.971875011920929, "step": 749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 241.0, "completions/max_terminated_length": 241.0, "completions/mean_length": 166.25, "completions/mean_terminated_length": 166.25, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.15, "grad_norm": 0.36535948514938354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8653308.0, "reward": 1.4062200784683228, "reward_std": 0.3749914765357971, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9374375343322754, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9937499761581421, "step": 750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 980.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 614.4375, "completions/mean_terminated_length": 614.4375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.1502, "grad_norm": 0.029649732634425163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8666707.0, "reward": 0.46249818801879883, "reward_std": 0.18195386230945587, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04999374598264694, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.908750057220459, "step": 751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 510.75, "completions/mean_terminated_length": 510.75, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.1504, "grad_norm": 0.04339483007788658, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8678975.0, "reward": 0.6996103525161743, "reward_std": 0.20404697954654694, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.39921873807907104, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6843750476837158, "step": 752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 154.0, "completions/max_terminated_length": 154.0, "completions/mean_length": 131.4375, "completions/mean_terminated_length": 131.4375, "completions/min_length": 95.0, "completions/min_terminated_length": 95.0, "epoch": 0.1506, "grad_norm": 1.4683764675282873e-05, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8684710.0, "reward": 1.4999732971191406, "reward_std": 2.60671313299099e-05, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.999943733215332, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9943749904632568, "step": 753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 438.3125, "completions/mean_terminated_length": 438.3125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.1508, "grad_norm": 0.11475616693496704, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8695627.0, "reward": 1.1056667566299438, "reward_std": 0.5404636859893799, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6488312482833862, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 659.875, "completions/mean_terminated_length": 635.6000366210938, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.151, "grad_norm": 0.07550064474344254, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8710009.0, "reward": 0.9000009894371033, "reward_std": 0.5074178576469421, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.48749998211860657, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.862500011920929, "step": 755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 313.625, "completions/mean_terminated_length": 313.625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1512, "grad_norm": 0.0001464469823986292, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8719395.0, "reward": 1.499232292175293, "reward_std": 0.000566332193557173, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9984624981880188, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 321.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 237.0625, "completions/mean_terminated_length": 237.0625, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.1514, "grad_norm": 0.0736432746052742, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8727188.0, "reward": 1.437935471534729, "reward_std": 0.2474628984928131, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9383687973022461, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9868749976158142, "step": 757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 389.9375, "completions/mean_terminated_length": 389.9375, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.1516, "grad_norm": 0.08938552439212799, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8738019.0, "reward": 1.0774917602539062, "reward_std": 0.45957672595977783, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.654981255531311, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8481250405311584, "step": 758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 540.4375, "completions/mean_terminated_length": 540.4375, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1518, "grad_norm": 0.10234562307596207, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8751146.0, "reward": 1.224435567855835, "reward_std": 0.5062791705131531, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7613687515258789, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 550.25, "completions/mean_terminated_length": 550.25, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.152, "grad_norm": 0.1150636151432991, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8763998.0, "reward": 0.668979287147522, "reward_std": 0.5350363850593567, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2754562497138977, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9781249761581421, "step": 760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 871.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 469.4375, "completions/mean_terminated_length": 469.4375, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.1522, "grad_norm": 0.07917208224534988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8775189.0, "reward": 0.6889071464538574, "reward_std": 0.3325551748275757, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3778125047683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.75, "step": 761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 590.875, "completions/mean_terminated_length": 562.0, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1524, "grad_norm": 0.03462425246834755, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8788467.0, "reward": 0.8617973923683167, "reward_std": 0.257607638835907, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7860937714576721, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.2718750238418579, "step": 762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 358.0, "completions/max_terminated_length": 358.0, "completions/mean_length": 247.8125, "completions/mean_terminated_length": 247.8125, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.1526, "grad_norm": 0.1487797647714615, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8796208.0, "reward": 0.8950573205947876, "reward_std": 0.4840486943721771, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4151124954223633, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 604.0, "completions/max_terminated_length": 604.0, "completions/mean_length": 490.4375, "completions/mean_terminated_length": 490.4375, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.1528, "grad_norm": 0.07729628682136536, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8807831.0, "reward": 1.2246696949005127, "reward_std": 0.5051106214523315, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7618374824523926, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 492.75, "completions/mean_terminated_length": 492.75, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.153, "grad_norm": 0.09831064194440842, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8819859.0, "reward": 1.0410165786743164, "reward_std": 0.4718541204929352, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.70703125, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.703125, "step": 765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 323.875, "completions/mean_terminated_length": 323.875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.1532, "grad_norm": 0.0834088921546936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8828961.0, "reward": 1.3751482963562012, "reward_std": 0.3396856188774109, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8752937316894531, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 447.6875, "completions/mean_terminated_length": 447.6875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.1534, "grad_norm": 0.04885704815387726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8839740.0, "reward": 0.8314069509506226, "reward_std": 0.2843826413154602, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5378124713897705, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.706250011920929, "step": 767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 830.75, "completions/mean_terminated_length": 786.1538696289062, "completions/min_length": 577.0, "completions/min_terminated_length": 577.0, "epoch": 0.1536, "grad_norm": 0.04697223752737045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8857688.0, "reward": 0.5930473804473877, "reward_std": 0.4331861436367035, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4985937476158142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.3531249761581421, "step": 768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 197.0, "completions/max_terminated_length": 197.0, "completions/mean_length": 142.5, "completions/mean_terminated_length": 142.5, "completions/min_length": 84.0, "completions/min_terminated_length": 84.0, "epoch": 0.1538, "grad_norm": 0.21214917302131653, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8863408.0, "reward": 0.7773886919021606, "reward_std": 0.43771883845329285, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3047749996185303, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 583.625, "completions/mean_terminated_length": 520.7142944335938, "completions/min_length": 460.0, "completions/min_terminated_length": 460.0, "epoch": 0.154, "grad_norm": 0.05215069279074669, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8876698.0, "reward": 0.539597749710083, "reward_std": 0.3167976140975952, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14169374108314514, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8331249952316284, "step": 770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 606.8125, "completions/mean_terminated_length": 579.0, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.1542, "grad_norm": 0.07365366071462631, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8890471.0, "reward": 0.7206227779388428, "reward_std": 0.49960222840309143, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3162437379360199, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8306249976158142, "step": 771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 380.875, "completions/mean_terminated_length": 380.875, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.1544, "grad_norm": 0.06409359723329544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8900613.0, "reward": 0.700522780418396, "reward_std": 0.32856857776641846, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27604374289512634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 417.4375, "completions/mean_terminated_length": 417.4375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.1546, "grad_norm": 0.09157534688711166, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8911676.0, "reward": 0.8991009593009949, "reward_std": 0.48018282651901245, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42320001125335693, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 309.1875, "completions/mean_terminated_length": 309.1875, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1548, "grad_norm": 0.08302626013755798, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8920415.0, "reward": 0.6967947483062744, "reward_std": 0.38096076250076294, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.33108749985694885, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7900000214576721, "step": 774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 432.75, "completions/mean_terminated_length": 432.75, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.155, "grad_norm": 0.09334935247898102, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8931947.0, "reward": 1.1368262767791748, "reward_std": 0.48367854952812195, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6486499905586243, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 826.0, "completions/max_terminated_length": 826.0, "completions/mean_length": 409.625, "completions/mean_terminated_length": 409.625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1552, "grad_norm": 0.11791979521512985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8942709.0, "reward": 0.943726122379303, "reward_std": 0.46444782614707947, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5124499797821045, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8849999904632568, "step": 776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 470.8125, "completions/mean_terminated_length": 470.8125, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.1554, "grad_norm": 0.07969141006469727, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8953890.0, "reward": 0.7151482105255127, "reward_std": 0.39168617129325867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2427937388420105, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 513.375, "completions/mean_terminated_length": 513.375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.1556, "grad_norm": 0.07082822173833847, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8966040.0, "reward": 1.3180856704711914, "reward_std": 0.3898948132991791, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8236687183380127, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 342.8125, "completions/mean_terminated_length": 342.8125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1558, "grad_norm": 0.14516226947307587, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8975509.0, "reward": 0.9540103673934937, "reward_std": 0.49745991826057434, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47051867842674255, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.979312539100647, "step": 779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 238.0, "completions/max_terminated_length": 238.0, "completions/mean_length": 187.125, "completions/mean_terminated_length": 187.125, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.156, "grad_norm": 0.08122260123491287, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8982039.0, "reward": 1.4377042055130005, "reward_std": 0.24740149080753326, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9379062652587891, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 435.75, "completions/mean_terminated_length": 396.5333557128906, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.1562, "grad_norm": 0.0749088004231453, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8993235.0, "reward": 0.7179697751998901, "reward_std": 0.38212519884109497, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3734375238418579, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.737500011920929, "step": 781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 964.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 567.75, "completions/mean_terminated_length": 567.75, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.1564, "grad_norm": 0.004956296179443598, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9006319.0, "reward": 0.5423449873924255, "reward_std": 0.02958306297659874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0846875011920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 783.5625, "completions/mean_terminated_length": 767.5333862304688, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.1566, "grad_norm": 0.09039270132780075, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9023048.0, "reward": 0.830929160118103, "reward_std": 0.5630998611450195, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4118562638759613, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 472.6875, "completions/mean_terminated_length": 435.933349609375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1568, "grad_norm": 0.09560110419988632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9035235.0, "reward": 0.5687322616577148, "reward_std": 0.39081650972366333, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2624624967575073, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.703125, "step": 784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 366.25, "completions/mean_terminated_length": 366.25, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.157, "grad_norm": 0.07309473305940628, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9045223.0, "reward": 0.6630351543426514, "reward_std": 0.27193692326545715, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3885687589645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6893749833106995, "step": 785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 789.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 557.625, "completions/mean_terminated_length": 557.625, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.1572, "grad_norm": 0.07536784559488297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9059873.0, "reward": 0.903425931930542, "reward_std": 0.47091639041900635, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43185001611709595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 647.4375, "completions/mean_terminated_length": 593.6428833007812, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.1574, "grad_norm": 0.035472698509693146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9074136.0, "reward": 0.40242287516593933, "reward_std": 0.24070008099079132, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05484374612569809, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 895.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 472.875, "completions/mean_terminated_length": 472.875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1576, "grad_norm": 0.031206564977765083, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9086022.0, "reward": 0.5705478191375732, "reward_std": 0.2154514342546463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20359373092651367, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.778124988079071, "step": 788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 789.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 574.1875, "completions/mean_terminated_length": 574.1875, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.1578, "grad_norm": 0.014113186858594418, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9099801.0, "reward": 0.5666228532791138, "reward_std": 0.08310282975435257, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13324373960494995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256250262260437, "step": 789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 396.9375, "completions/mean_terminated_length": 396.9375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.158, "grad_norm": 0.12326539307832718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9110872.0, "reward": 0.9860977530479431, "reward_std": 0.5459200143814087, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.534693717956543, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 153.125, "completions/mean_terminated_length": 153.125, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "epoch": 0.1582, "grad_norm": 0.0881800726056099, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9116986.0, "reward": 0.5637456178665161, "reward_std": 0.24969126284122467, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06498749554157257, "rewards/confidence_one_or_zero": 0.875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9987499713897705, "step": 791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 590.375, "completions/mean_terminated_length": 561.4666748046875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.1584, "grad_norm": 0.10174448043107986, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9130528.0, "reward": 0.6348320245742798, "reward_std": 0.5110194683074951, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3321624994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 600.875, "completions/mean_terminated_length": 503.23077392578125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.1586, "grad_norm": 0.11942103505134583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9144494.0, "reward": 0.6965633630752563, "reward_std": 0.6238778233528137, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3931249976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.721875011920929, "step": 793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1007.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 723.4375, "completions/mean_terminated_length": 723.4375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.1588, "grad_norm": 0.04361611232161522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9160165.0, "reward": 0.6596041321754456, "reward_std": 0.3280012607574463, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19420623779296875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 367.8125, "completions/mean_terminated_length": 367.8125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.159, "grad_norm": 0.08972814679145813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9170514.0, "reward": 0.967332124710083, "reward_std": 0.4843050241470337, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49716246128082275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1022.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 605.8125, "completions/mean_terminated_length": 605.8125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.1592, "grad_norm": 0.06643634289503098, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9184719.0, "reward": 0.6335134506225586, "reward_std": 0.3650873005390167, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20452500879764557, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 344.0625, "completions/mean_terminated_length": 344.0625, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.1594, "grad_norm": 0.04054255411028862, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9194304.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 542.1875, "completions/mean_terminated_length": 542.1875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1596, "grad_norm": 0.07628677785396576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9207043.0, "reward": 1.0522416830062866, "reward_std": 0.4744350016117096, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6044812798500061, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 364.1875, "completions/mean_terminated_length": 364.1875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.1598, "grad_norm": 0.00016927153046708554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9216646.0, "reward": 1.4993764162063599, "reward_std": 0.0006459736614488065, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987499713897705, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 494.5625, "completions/mean_terminated_length": 494.5625, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.16, "grad_norm": 0.03847755864262581, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9228815.0, "reward": 0.7842192649841309, "reward_std": 0.20043246448040009, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5684374570846558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5499999523162842, "step": 800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 276.0, "completions/max_terminated_length": 276.0, "completions/mean_length": 240.75, "completions/mean_terminated_length": 240.75, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1602, "grad_norm": 0.05718431621789932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9236459.0, "reward": 0.5917043685913086, "reward_std": 0.24287572503089905, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12090624868869781, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 656.25, "completions/mean_terminated_length": 631.7333374023438, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.1604, "grad_norm": 0.0554158091545105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9250991.0, "reward": 0.7489042282104492, "reward_std": 0.4677391052246094, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31030625104904175, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 324.9375, "completions/mean_terminated_length": 324.9375, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1606, "grad_norm": 0.0932435691356659, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9259982.0, "reward": 0.629585325717926, "reward_std": 0.36544543504714966, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19666874408721924, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.956250011920929, "step": 803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 285.125, "completions/mean_terminated_length": 285.125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.1608, "grad_norm": 0.08566929399967194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9268432.0, "reward": 1.380285382270813, "reward_std": 0.32459819316864014, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855687379837036, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 539.625, "completions/mean_terminated_length": 507.3333740234375, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.161, "grad_norm": 0.045496534556150436, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9281114.0, "reward": 0.560547947883606, "reward_std": 0.28245365619659424, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24609375, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7406249642372131, "step": 805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 620.25, "completions/mean_terminated_length": 620.25, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.1612, "grad_norm": 0.06211080029606819, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9295582.0, "reward": 0.3892202079296112, "reward_std": 0.2790306806564331, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09093749523162842, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9468750357627869, "step": 806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 888.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 392.6875, "completions/mean_terminated_length": 392.6875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.1614, "grad_norm": 0.0761452317237854, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9305961.0, "reward": 0.5629292130470276, "reward_std": 0.28364354372024536, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12585625052452087, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 767.5, "completions/mean_terminated_length": 682.0, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.1616, "grad_norm": 0.05014438182115555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9322865.0, "reward": 0.3831912875175476, "reward_std": 0.3815658688545227, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20388123393058777, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4925000071525574, "step": 808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1010.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 658.0, "completions/mean_terminated_length": 658.0, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.1618, "grad_norm": 0.041022997349500656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9338161.0, "reward": 0.9123443365097046, "reward_std": 0.30345460772514343, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6996874809265137, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.47499996423721313, "step": 809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 363.5625, "completions/mean_terminated_length": 363.5625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.162, "grad_norm": 0.13824012875556946, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9347802.0, "reward": 0.7374199628829956, "reward_std": 0.506950318813324, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3498374819755554, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 682.75, "completions/mean_terminated_length": 660.0000610351562, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.1622, "grad_norm": 0.07042676955461502, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9365078.0, "reward": 0.5028011202812195, "reward_std": 0.46835198998451233, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1931000053882599, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 766.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 547.8125, "completions/mean_terminated_length": 547.8125, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.1624, "grad_norm": 0.08146604895591736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9378131.0, "reward": 0.906482458114624, "reward_std": 0.4777293801307678, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43796247243881226, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 333.4375, "completions/mean_terminated_length": 333.4375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.1626, "grad_norm": 0.06041460111737251, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9387018.0, "reward": 1.439450979232788, "reward_std": 0.23752017319202423, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413999915122986, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 778.75, "completions/mean_terminated_length": 697.0, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.1628, "grad_norm": 0.045534174889326096, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9403558.0, "reward": 0.4185161888599396, "reward_std": 0.35915663838386536, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21203124523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5750000476837158, "step": 814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 581.875, "completions/mean_terminated_length": 518.7142944335938, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.163, "grad_norm": 0.13604849576950073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9416948.0, "reward": 1.0970319509506226, "reward_std": 0.5472418069839478, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7565624713897705, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.675000011920929, "step": 815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 246.5, "completions/mean_terminated_length": 246.5, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.1632, "grad_norm": 0.14168241620063782, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9424908.0, "reward": 0.7864041328430176, "reward_std": 0.4251270592212677, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3228062391281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 535.9375, "completions/mean_terminated_length": 535.9375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.1634, "grad_norm": 0.06721587479114532, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9437035.0, "reward": 0.7074883580207825, "reward_std": 0.32668977975845337, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28997498750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 609.25, "completions/mean_terminated_length": 609.25, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.1636, "grad_norm": 0.09744866192340851, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9450799.0, "reward": 0.7065633535385132, "reward_std": 0.46309494972229004, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.35062500834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 483.0, "completions/max_terminated_length": 483.0, "completions/mean_length": 348.625, "completions/mean_terminated_length": 348.625, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.1638, "grad_norm": 0.1341971904039383, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9460041.0, "reward": 0.8269293308258057, "reward_std": 0.4684930443763733, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34135621786117554, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 517.4375, "completions/mean_terminated_length": 483.66668701171875, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.164, "grad_norm": 0.058006614446640015, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9472352.0, "reward": 0.7084380388259888, "reward_std": 0.3324069380760193, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5418750047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.503125011920929, "step": 820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 321.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 219.3125, "completions/mean_terminated_length": 219.3125, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.1642, "grad_norm": 0.07109778374433517, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9479733.0, "reward": 0.5814637541770935, "reward_std": 0.24579095840454102, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10042500495910645, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 585.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 343.75, "completions/mean_terminated_length": 343.75, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.1644, "grad_norm": 0.0023844956886023283, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9489009.0, "reward": 0.5445165634155273, "reward_std": 0.011709420010447502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 225.0, "completions/max_terminated_length": 225.0, "completions/mean_length": 177.0625, "completions/mean_terminated_length": 177.0625, "completions/min_length": 137.0, "completions/min_terminated_length": 137.0, "epoch": 0.1646, "grad_norm": 0.10799860209226608, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9495346.0, "reward": 1.002501130104065, "reward_std": 0.3017723858356476, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6299999952316284, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3499999940395355, "step": 823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 281.0, "completions/max_terminated_length": 281.0, "completions/mean_length": 244.125, "completions/mean_terminated_length": 244.125, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.1648, "grad_norm": 0.15549229085445404, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9503412.0, "reward": 1.0784165859222412, "reward_std": 0.49268361926078796, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5943312644958496, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 246.0, "completions/max_terminated_length": 246.0, "completions/mean_length": 152.625, "completions/mean_terminated_length": 152.625, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "epoch": 0.165, "grad_norm": 0.23183514177799225, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9509502.0, "reward": 1.003554344177246, "reward_std": 0.5123699307441711, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5071062445640564, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.986875057220459, "step": 825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 205.5, "completions/mean_terminated_length": 205.5, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.1652, "grad_norm": 0.18617261946201324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9516806.0, "reward": 0.8341760039329529, "reward_std": 0.46318319439888, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3558500111103058, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 360.0, "completions/max_terminated_length": 360.0, "completions/mean_length": 277.625, "completions/mean_terminated_length": 277.625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1654, "grad_norm": 0.15979507565498352, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9524992.0, "reward": 1.2928698062896729, "reward_std": 0.38182303309440613, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8357374668121338, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 276.0625, "completions/mean_terminated_length": 276.0625, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.1656, "grad_norm": 0.10545014590024948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9533825.0, "reward": 0.7129292488098145, "reward_std": 0.39029908180236816, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23835626244544983, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 316.25, "completions/mean_terminated_length": 316.25, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.1658, "grad_norm": 0.10642413794994354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9542933.0, "reward": 1.3178728818893433, "reward_std": 0.38978931307792664, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8232437968254089, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 288.9375, "completions/mean_terminated_length": 288.9375, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.166, "grad_norm": 0.053273677825927734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9551460.0, "reward": 0.7065632343292236, "reward_std": 0.20346103608608246, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4131249785423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 324.25, "completions/mean_terminated_length": 324.25, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.1662, "grad_norm": 0.10128859430551529, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9561048.0, "reward": 0.8401541113853455, "reward_std": 0.4588193893432617, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36780625581741333, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 838.875, "completions/mean_terminated_length": 727.7999877929688, "completions/min_length": 546.0, "completions/min_terminated_length": 546.0, "epoch": 0.1664, "grad_norm": 0.0974939614534378, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9580246.0, "reward": 0.5441409945487976, "reward_std": 0.5551596283912659, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.40078121423721313, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4000000059604645, "step": 832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 661.75, "completions/mean_terminated_length": 661.75, "completions/min_length": 538.0, "completions/min_terminated_length": 538.0, "epoch": 0.1666, "grad_norm": 0.029579492285847664, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9595890.0, "reward": 0.6790633201599121, "reward_std": 0.1668575555086136, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3581250011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 420.75, "completions/mean_terminated_length": 420.75, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.1668, "grad_norm": 0.11036846041679382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9606414.0, "reward": 0.9346884489059448, "reward_std": 0.5257036685943604, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49437499046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9187500476837158, "step": 834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 755.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 540.0, "completions/mean_terminated_length": 540.0, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.167, "grad_norm": 0.08322430402040482, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9618958.0, "reward": 1.2797480821609497, "reward_std": 0.407776415348053, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.9344937205314636, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.75, "step": 835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 364.625, "completions/mean_terminated_length": 364.625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1672, "grad_norm": 0.07916955649852753, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9629208.0, "reward": 0.7269541025161743, "reward_std": 0.3831254243850708, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26640623807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 537.0, "completions/max_terminated_length": 537.0, "completions/mean_length": 420.875, "completions/mean_terminated_length": 420.875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.1674, "grad_norm": 0.08680418878793716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9640454.0, "reward": 1.2864696979522705, "reward_std": 0.4711858034133911, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8229374885559082, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9512499570846558, "step": 837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 388.0, "completions/max_terminated_length": 388.0, "completions/mean_length": 257.9375, "completions/mean_terminated_length": 257.9375, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.1676, "grad_norm": 0.1002885177731514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9648533.0, "reward": 1.3186509609222412, "reward_std": 0.38848060369491577, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8248000144958496, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 548.375, "completions/mean_terminated_length": 516.6666870117188, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.1678, "grad_norm": 0.07704009860754013, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9661595.0, "reward": 1.151841640472412, "reward_std": 0.523673951625824, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6786812543869019, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 269.0, "completions/max_terminated_length": 269.0, "completions/mean_length": 216.25, "completions/mean_terminated_length": 216.25, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.168, "grad_norm": 0.05976245552301407, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9669135.0, "reward": 1.439582109451294, "reward_std": 0.2375553995370865, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416624903678894, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 271.0, "completions/max_terminated_length": 271.0, "completions/mean_length": 212.9375, "completions/mean_terminated_length": 212.9375, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.1682, "grad_norm": 0.14533331990242004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9676334.0, "reward": 1.0463167428970337, "reward_std": 0.5459311604499817, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5926312208175659, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 440.0625, "completions/mean_terminated_length": 440.0625, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1684, "grad_norm": 0.08474050462245941, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9687503.0, "reward": 0.8395822048187256, "reward_std": 0.45915547013282776, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36666250228881836, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 483.9375, "completions/mean_terminated_length": 483.9375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1686, "grad_norm": 0.010570251382887363, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9698878.0, "reward": 0.5709384083747864, "reward_std": 0.04535386711359024, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1418749988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.925000011920929, "step": 843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 491.0, "completions/max_terminated_length": 491.0, "completions/mean_length": 367.125, "completions/mean_terminated_length": 367.125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.1688, "grad_norm": 0.04807748645544052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9708432.0, "reward": 1.4399478435516357, "reward_std": 0.2376532107591629, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9423937797546387, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 275.4375, "completions/mean_terminated_length": 275.4375, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.169, "grad_norm": 0.12600071728229523, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9716551.0, "reward": 0.8343417644500732, "reward_std": 0.463437020778656, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.356181263923645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 803.125, "completions/mean_terminated_length": 670.6000366210938, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.1692, "grad_norm": 0.04355628415942192, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9733897.0, "reward": 0.26281020045280457, "reward_std": 0.27196311950683594, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.025618748739361763, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.5181249976158142, "step": 846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 419.5625, "completions/mean_terminated_length": 379.2666931152344, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.1694, "grad_norm": 0.10523470491170883, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9744210.0, "reward": 1.104219675064087, "reward_std": 0.5393410325050354, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.645937442779541, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8812499642372131, "step": 847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 512.0625, "completions/mean_terminated_length": 512.0625, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.1696, "grad_norm": 0.0977347269654274, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9755971.0, "reward": 1.1318261623382568, "reward_std": 0.4900140166282654, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6386500000953674, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 205.5, "completions/mean_terminated_length": 205.5, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 0.1698, "grad_norm": 0.15123388171195984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9763067.0, "reward": 1.2845263481140137, "reward_std": 0.39935049414634705, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8190500140190125, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 576.5, "completions/mean_terminated_length": 512.5714721679688, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.17, "grad_norm": 0.11512955278158188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9775971.0, "reward": 0.587422788143158, "reward_std": 0.4300609529018402, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23734374344348907, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 351.0, "completions/max_terminated_length": 351.0, "completions/mean_length": 278.125, "completions/mean_terminated_length": 278.125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.1702, "grad_norm": 0.15830640494823456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9784197.0, "reward": 1.009101152420044, "reward_std": 0.5069533586502075, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5182000398635864, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 780.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 541.125, "completions/mean_terminated_length": 541.125, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.1704, "grad_norm": 0.1204637661576271, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9798103.0, "reward": 0.9269415140151978, "reward_std": 0.5679826140403748, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5413812398910522, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 241.6875, "completions/mean_terminated_length": 241.6875, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.1706, "grad_norm": 0.12556494772434235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9805426.0, "reward": 1.321294903755188, "reward_std": 0.38328951597213745, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8300875425338745, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 731.4375, "completions/mean_terminated_length": 711.933349609375, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.1708, "grad_norm": 0.042470771819353104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9821609.0, "reward": 0.6791290044784546, "reward_std": 0.3721710443496704, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2957562208175659, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8268749713897705, "step": 854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 999.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 688.625, "completions/mean_terminated_length": 688.625, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.171, "grad_norm": 0.02391275018453598, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9836755.0, "reward": 0.6304446458816528, "reward_std": 0.14342810213565826, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2608875036239624, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8350000381469727, "step": 855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 407.0, "completions/max_terminated_length": 407.0, "completions/mean_length": 310.9375, "completions/mean_terminated_length": 310.9375, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.1712, "grad_norm": 0.11677910387516022, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9845538.0, "reward": 1.2614572048187256, "reward_std": 0.424976110458374, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7729125022888184, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 692.1875, "completions/mean_terminated_length": 581.5833740234375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.1714, "grad_norm": 0.07296936959028244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9860437.0, "reward": 0.580835223197937, "reward_std": 0.5387163758277893, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2866687476634979, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6818749904632568, "step": 857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 530.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 347.5, "completions/mean_terminated_length": 347.5, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.1716, "grad_norm": 0.1503760665655136, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9869645.0, "reward": 0.9715637564659119, "reward_std": 0.48270919919013977, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5056250095367432, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 386.4375, "completions/mean_terminated_length": 386.4375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.1718, "grad_norm": 0.019754579290747643, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9882756.0, "reward": 0.5086699724197388, "reward_std": 0.1386445164680481, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07983750104904175, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 291.9375, "completions/mean_terminated_length": 291.9375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.172, "grad_norm": 0.06528910249471664, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9890819.0, "reward": 1.4372262954711914, "reward_std": 0.24992693960666656, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9369499683380127, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 246.75, "completions/mean_terminated_length": 246.75, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1722, "grad_norm": 0.007812474388629198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9898303.0, "reward": 0.5239667892456055, "reward_std": 0.018698127940297127, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.047931186854839325, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9755624532699585, "step": 861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 810.625, "completions/mean_terminated_length": 780.1428833007812, "completions/min_length": 591.0, "completions/min_terminated_length": 591.0, "epoch": 0.1724, "grad_norm": 0.07683074474334717, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9917001.0, "reward": 0.598125696182251, "reward_std": 0.5329092144966125, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3212500214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 571.4375, "completions/mean_terminated_length": 571.4375, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.1726, "grad_norm": 0.09045020490884781, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9929664.0, "reward": 0.9421699047088623, "reward_std": 0.5330116748809814, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5093374848365784, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9412500262260437, "step": 863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 454.0, "completions/max_terminated_length": 454.0, "completions/mean_length": 337.0625, "completions/mean_terminated_length": 337.0625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1728, "grad_norm": 0.13783778250217438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9938721.0, "reward": 0.9079697132110596, "reward_std": 0.4729275405406952, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44093748927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 550.0, "completions/mean_terminated_length": 550.0, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.173, "grad_norm": 0.0802881196141243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9951217.0, "reward": 0.6781418323516846, "reward_std": 0.4285159707069397, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23128125071525574, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.971875011920929, "step": 865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 972.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 390.0625, "completions/mean_terminated_length": 390.0625, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.1732, "grad_norm": 0.14848843216896057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9961378.0, "reward": 0.8041260838508606, "reward_std": 0.5018141269683838, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3582500219345093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 601.125, "completions/mean_terminated_length": 503.5384826660156, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1734, "grad_norm": 0.061219483613967896, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9974884.0, "reward": 0.6863163709640503, "reward_std": 0.530077338218689, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3101312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7706249952316284, "step": 867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 524.0625, "completions/mean_terminated_length": 524.0625, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.1736, "grad_norm": 0.07315431535243988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9987445.0, "reward": 1.0753542184829712, "reward_std": 0.4963269531726837, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5882062315940857, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 862.625, "completions/mean_terminated_length": 789.2727661132812, "completions/min_length": 608.0, "completions/min_terminated_length": 608.0, "epoch": 0.1738, "grad_norm": 0.06388141959905624, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10006207.0, "reward": 0.6293631792068481, "reward_std": 0.5533629655838013, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.38372498750686646, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5299999713897705, "step": 869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 347.25, "completions/mean_terminated_length": 347.25, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.174, "grad_norm": 0.16815796494483948, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10015795.0, "reward": 1.1315979957580566, "reward_std": 0.5841472148895264, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7006937265396118, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 733.8125, "completions/mean_terminated_length": 714.4666748046875, "completions/min_length": 553.0, "completions/min_terminated_length": 553.0, "epoch": 0.1742, "grad_norm": 0.07920420169830322, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10031920.0, "reward": 1.0354260206222534, "reward_std": 0.6409369111061096, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6333500146865845, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9187500476837158, "step": 871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 524.3125, "completions/mean_terminated_length": 524.3125, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.1744, "grad_norm": 0.07659240067005157, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10044661.0, "reward": 0.6436728239059448, "reward_std": 0.36472630500793457, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22484374046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9399999380111694, "step": 872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 475.5, "completions/mean_terminated_length": 397.14288330078125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1746, "grad_norm": 0.09523595124483109, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10055917.0, "reward": 0.7067822813987732, "reward_std": 0.5048380494117737, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28856247663497925, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8450000286102295, "step": 873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 355.125, "completions/mean_terminated_length": 355.125, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.1748, "grad_norm": 0.1638733595609665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10065775.0, "reward": 0.8740478754043579, "reward_std": 0.44942131638526917, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.43559375405311584, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 396.75, "completions/mean_terminated_length": 396.75, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.175, "grad_norm": 0.05982143059372902, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10075931.0, "reward": 0.7091412544250488, "reward_std": 0.23852556943893433, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4807812571525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 332.0625, "completions/mean_terminated_length": 332.0625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1752, "grad_norm": 0.003575744107365608, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10085852.0, "reward": 0.5420916080474854, "reward_std": 0.01446700282394886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08418124914169312, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 627.25, "completions/mean_terminated_length": 600.800048828125, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.1754, "grad_norm": 0.0574130155146122, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10100960.0, "reward": 0.6731914281845093, "reward_std": 0.484186053276062, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3463812470436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7756249904632568, "step": 877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 338.375, "completions/mean_terminated_length": 338.375, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.1756, "grad_norm": 0.11777281761169434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10110150.0, "reward": 1.136294960975647, "reward_std": 0.4844037890434265, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6475875377655029, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 404.875, "completions/mean_terminated_length": 404.875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1758, "grad_norm": 0.09484431147575378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10120276.0, "reward": 1.3190228939056396, "reward_std": 0.38699793815612793, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8255437612533569, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 623.4375, "completions/mean_terminated_length": 623.4375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.176, "grad_norm": 0.052552927285432816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10134155.0, "reward": 0.7255412340164185, "reward_std": 0.28427770733833313, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5135812759399414, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5362499952316284, "step": 880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 362.625, "completions/mean_terminated_length": 362.625, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1762, "grad_norm": 0.0023804036900401115, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10143925.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 459.375, "completions/mean_terminated_length": 459.375, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.1764, "grad_norm": 0.050574544817209244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10155259.0, "reward": 0.6644541621208191, "reward_std": 0.32590189576148987, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20390623807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 618.875, "completions/mean_terminated_length": 618.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1766, "grad_norm": 0.048829421401023865, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10169017.0, "reward": 0.8122416734695435, "reward_std": 0.3799396753311157, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4369812309741974, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8006249666213989, "step": 883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 392.375, "completions/mean_terminated_length": 392.375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.1768, "grad_norm": 0.10646852105855942, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10179967.0, "reward": 1.1953635215759277, "reward_std": 0.46522295475006104, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7032250165939331, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 408.6875, "completions/mean_terminated_length": 367.66668701171875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.177, "grad_norm": 0.09934016317129135, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10190458.0, "reward": 0.686069905757904, "reward_std": 0.42549827694892883, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24713748693466187, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 720.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 505.3125, "completions/mean_terminated_length": 505.3125, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.1772, "grad_norm": 0.0732155367732048, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10202607.0, "reward": 0.7885946035385132, "reward_std": 0.4263220429420471, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.45218750834465027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 404.625, "completions/mean_terminated_length": 404.625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.1774, "grad_norm": 0.05367175489664078, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10212873.0, "reward": 0.6540601253509521, "reward_std": 0.25077688694000244, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3706187605857849, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6681250333786011, "step": 887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 958.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 726.625, "completions/mean_terminated_length": 726.625, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.1776, "grad_norm": 0.005666845012456179, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10228323.0, "reward": 0.5537511110305786, "reward_std": 0.03976471722126007, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10750000178813934, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437499642372131, "step": 888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 615.75, "completions/mean_terminated_length": 588.5333862304688, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.1778, "grad_norm": 0.0644797831773758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10242271.0, "reward": 0.7520978450775146, "reward_std": 0.4299311935901642, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3791937232017517, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7768750190734863, "step": 889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 517.4375, "completions/mean_terminated_length": 517.4375, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.178, "grad_norm": 0.08692436665296555, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10254646.0, "reward": 0.940285861492157, "reward_std": 0.5411327481269836, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5055687427520752, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.921875, "step": 890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 343.5, "completions/mean_terminated_length": 343.5, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.1782, "grad_norm": 0.008401608094573021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10263854.0, "reward": 0.5489041805267334, "reward_std": 0.02906905859708786, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09780624508857727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 713.5, "completions/mean_terminated_length": 669.1428833007812, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.1784, "grad_norm": 0.08645788580179214, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10280790.0, "reward": 0.6803663969039917, "reward_std": 0.5609583258628845, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.36073124408721924, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7037500143051147, "step": 892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 584.125, "completions/mean_terminated_length": 482.61541748046875, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1786, "grad_norm": 0.05545772984623909, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10294408.0, "reward": 0.49914154410362244, "reward_std": 0.34363335371017456, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12328124791383743, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.778124988079071, "step": 893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 440.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 296.0625, "completions/mean_terminated_length": 296.0625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.1788, "grad_norm": 0.13877634704113007, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10303401.0, "reward": 0.8962104320526123, "reward_std": 0.4828832149505615, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4174187481403351, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 481.4375, "completions/mean_terminated_length": 481.4375, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.179, "grad_norm": 0.10367999225854874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10314944.0, "reward": 1.143669605255127, "reward_std": 0.45824721455574036, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7248375415802002, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6431249976158142, "step": 895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 666.5, "completions/mean_terminated_length": 666.5, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.1792, "grad_norm": 0.07944636046886444, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10329624.0, "reward": 1.0350792407989502, "reward_std": 0.47928985953330994, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5701562762260437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 991.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 706.3125, "completions/mean_terminated_length": 706.3125, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.1794, "grad_norm": 0.03321519121527672, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10345133.0, "reward": 0.6572539806365967, "reward_std": 0.25774049758911133, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2520062327384949, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 514.5, "completions/mean_terminated_length": 514.5, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.1796, "grad_norm": 0.003894299501553178, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10356885.0, "reward": 0.5431230068206787, "reward_std": 0.02293558046221733, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08624374866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249380111694, "step": 898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 273.625, "completions/mean_terminated_length": 273.625, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.1798, "grad_norm": 0.1792493462562561, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10364959.0, "reward": 1.1995165348052979, "reward_std": 0.45862603187561035, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7115312814712524, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 679.375, "completions/mean_terminated_length": 679.375, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.18, "grad_norm": 0.034456077963113785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10379637.0, "reward": 0.6454447507858276, "reward_std": 0.25887730717658997, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1658874899148941, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8787499666213989, "step": 900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 386.0625, "completions/mean_terminated_length": 386.0625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.1802, "grad_norm": 0.10171154141426086, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10390118.0, "reward": 1.0463230609893799, "reward_std": 0.4810199439525604, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5926437377929688, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 338.0, "completions/max_terminated_length": 338.0, "completions/mean_length": 279.3125, "completions/mean_terminated_length": 279.3125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1804, "grad_norm": 0.06682498008012772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10398171.0, "reward": 1.4395947456359863, "reward_std": 0.23755881190299988, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416874647140503, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 398.5625, "completions/mean_terminated_length": 398.5625, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.1806, "grad_norm": 0.003465190762653947, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10409300.0, "reward": 0.5360416173934937, "reward_std": 0.017606575042009354, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07208124548196793, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 551.3125, "completions/mean_terminated_length": 551.3125, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.1808, "grad_norm": 0.11576033383607864, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10422345.0, "reward": 0.9215604066848755, "reward_std": 0.5416698455810547, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4681187570095062, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 353.8125, "completions/mean_terminated_length": 353.8125, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.181, "grad_norm": 0.10559242218732834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10431798.0, "reward": 1.093425989151001, "reward_std": 0.4353608787059784, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6868500113487244, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8112500309944153, "step": 905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 645.625, "completions/mean_terminated_length": 591.5714721679688, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.1812, "grad_norm": 0.08092433214187622, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10446672.0, "reward": 0.658969521522522, "reward_std": 0.525209367275238, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3804374933242798, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6668750047683716, "step": 906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 533.375, "completions/mean_terminated_length": 500.66668701171875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1814, "grad_norm": 0.04710749164223671, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10459478.0, "reward": 0.752175509929657, "reward_std": 0.2884870171546936, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5668500065803528, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.45499998331069946, "step": 907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 329.3125, "completions/mean_terminated_length": 329.3125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.1816, "grad_norm": 0.09821297973394394, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10469883.0, "reward": 1.0154728889465332, "reward_std": 0.4997072219848633, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.530943751335144, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 242.0, "completions/max_terminated_length": 242.0, "completions/mean_length": 173.125, "completions/mean_terminated_length": 173.125, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.1818, "grad_norm": 0.13460740447044373, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10476621.0, "reward": 1.4384386539459229, "reward_std": 0.24497050046920776, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9393750429153442, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "step": 909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 691.25, "completions/mean_terminated_length": 669.0667114257812, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.182, "grad_norm": 0.08585113286972046, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10492497.0, "reward": 0.99062579870224, "reward_std": 0.5385201573371887, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.668749988079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7493749856948853, "step": 910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 495.9375, "completions/mean_terminated_length": 495.9375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.1822, "grad_norm": 0.06075114384293556, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10504368.0, "reward": 0.6165603995323181, "reward_std": 0.2461388111114502, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17061874270439148, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9368749856948853, "step": 911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 280.3125, "completions/mean_terminated_length": 280.3125, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1824, "grad_norm": 0.16189807653427124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10512789.0, "reward": 1.1725759506225586, "reward_std": 0.44673851132392883, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7201499938964844, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9087499976158142, "step": 912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 342.0, "completions/mean_terminated_length": 342.0, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.1826, "grad_norm": 0.1183106079697609, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10522485.0, "reward": 0.9085071682929993, "reward_std": 0.4729026257991791, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4420124888420105, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 666.125, "completions/mean_terminated_length": 615.0, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.1828, "grad_norm": 0.07218990474939346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10537063.0, "reward": 0.6721887588500977, "reward_std": 0.4665946066379547, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.28187501430511475, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 341.25, "completions/mean_terminated_length": 341.25, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.183, "grad_norm": 0.12906058132648468, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10546459.0, "reward": 0.982810378074646, "reward_std": 0.4827153980731964, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5281187295913696, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9118749499320984, "step": 915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 465.0, "completions/mean_terminated_length": 465.0, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.1832, "grad_norm": 0.08136321604251862, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10558427.0, "reward": 0.7261295318603516, "reward_std": 0.47854045033454895, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2647562325000763, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9900000095367432, "step": 916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 633.375, "completions/mean_terminated_length": 607.3333740234375, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.1834, "grad_norm": 0.0709771141409874, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10573473.0, "reward": 0.6211696863174438, "reward_std": 0.4107617139816284, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24233749508857727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8431249856948853, "step": 917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 414.1875, "completions/mean_terminated_length": 414.1875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.1836, "grad_norm": 0.09917768090963364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10584020.0, "reward": 0.8557823896408081, "reward_std": 0.44999298453330994, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39906251430511475, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 338.125, "completions/mean_terminated_length": 338.125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.1838, "grad_norm": 0.1649937629699707, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10593238.0, "reward": 1.199747920036316, "reward_std": 0.4587872624397278, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7119937539100647, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 793.625, "completions/mean_terminated_length": 716.8333740234375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.184, "grad_norm": 0.06999846547842026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10609712.0, "reward": 0.492188036441803, "reward_std": 0.4600023031234741, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.296875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.48749998211860657, "step": 920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 382.5, "completions/mean_terminated_length": 382.5, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.1842, "grad_norm": 0.101331427693367, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10619912.0, "reward": 1.0731542110443115, "reward_std": 0.49936309456825256, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5838062167167664, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 295.875, "completions/mean_terminated_length": 295.875, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.1844, "grad_norm": 0.14621728658676147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10628902.0, "reward": 0.9582823514938354, "reward_std": 0.4925306737422943, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4790624976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 309.625, "completions/mean_terminated_length": 309.625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.1846, "grad_norm": 0.14906661212444305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10638352.0, "reward": 1.2253729104995728, "reward_std": 0.5042374134063721, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7632437944412231, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 384.0, "completions/max_terminated_length": 384.0, "completions/mean_length": 243.5, "completions/mean_terminated_length": 243.5, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.1848, "grad_norm": 0.09108202904462814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10646120.0, "reward": 0.6571729779243469, "reward_std": 0.3290401101112366, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18934375047683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 388.5625, "completions/mean_terminated_length": 388.5625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.185, "grad_norm": 0.09955795109272003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10656529.0, "reward": 0.8447670936584473, "reward_std": 0.46446698904037476, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37703123688697815, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 508.625, "completions/mean_terminated_length": 508.625, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.1852, "grad_norm": 0.2074775993824005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10669275.0, "reward": 0.896460235118866, "reward_std": 0.39328956604003906, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5429187417030334, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7381249666213989, "step": 926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 377.125, "completions/mean_terminated_length": 377.125, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.1854, "grad_norm": 0.08646026998758316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10679277.0, "reward": 0.6332041025161743, "reward_std": 0.3644181787967682, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20390623807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 714.8125, "completions/mean_terminated_length": 670.6428833007812, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.1856, "grad_norm": 0.0832647755742073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10694874.0, "reward": 0.5682045221328735, "reward_std": 0.4410482943058014, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19890624284744263, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.734375, "step": 928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 588.625, "completions/mean_terminated_length": 559.6000366210938, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.1858, "grad_norm": 0.05391903966665268, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10707716.0, "reward": 0.5340603590011597, "reward_std": 0.3167736232280731, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13061875104904175, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 800.0, "completions/mean_terminated_length": 748.3077392578125, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.186, "grad_norm": 0.07327055186033249, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10724676.0, "reward": 0.5189069509506226, "reward_std": 0.5068359971046448, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2878125011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.684374988079071, "step": 930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 843.4375, "completions/mean_terminated_length": 801.769287109375, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.1862, "grad_norm": 0.035083282738924026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10743739.0, "reward": 0.4613163471221924, "reward_std": 0.3567916750907898, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23513124883174896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5887500047683716, "step": 931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 515.9375, "completions/mean_terminated_length": 443.357177734375, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.1864, "grad_norm": 0.10026230663061142, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10755914.0, "reward": 1.1460946798324585, "reward_std": 0.569258451461792, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7296874523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 480.1875, "completions/mean_terminated_length": 480.1875, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.1866, "grad_norm": 0.09158847481012344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10767853.0, "reward": 0.8371855020523071, "reward_std": 0.46115973591804504, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3618687391281128, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 778.9375, "completions/mean_terminated_length": 697.25, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.1868, "grad_norm": 0.05332690104842186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10784908.0, "reward": 0.5675787925720215, "reward_std": 0.4314081370830536, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3226562440395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.534375011920929, "step": 934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 317.5, "completions/mean_terminated_length": 317.5, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.187, "grad_norm": 0.13454414904117584, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10794292.0, "reward": 1.2829415798187256, "reward_std": 0.39340370893478394, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8158812522888184, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 722.5625, "completions/mean_terminated_length": 702.4666748046875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1872, "grad_norm": 0.07713615894317627, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10809869.0, "reward": 0.9179445505142212, "reward_std": 0.5221693515777588, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5858874917030334, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6568750143051147, "step": 936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 316.5, "completions/mean_terminated_length": 316.5, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1874, "grad_norm": 0.12639324367046356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10818885.0, "reward": 1.009760856628418, "reward_std": 0.4718018174171448, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5820187926292419, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8518750071525574, "step": 937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 584.6875, "completions/mean_terminated_length": 584.6875, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.1876, "grad_norm": 0.07937845587730408, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10832352.0, "reward": 0.5948448181152344, "reward_std": 0.43403056263923645, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25218749046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 637.4375, "completions/mean_terminated_length": 582.2142944335938, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.1878, "grad_norm": 0.04196765273809433, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10847175.0, "reward": 0.4104635417461395, "reward_std": 0.25182175636291504, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07092499732971191, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7706249952316284, "step": 939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 485.25, "completions/mean_terminated_length": 485.25, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.188, "grad_norm": 0.08646245300769806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10858955.0, "reward": 0.7215572595596313, "reward_std": 0.386285662651062, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25561249256134033, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 368.625, "completions/mean_terminated_length": 368.625, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.1882, "grad_norm": 0.18191616237163544, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10869605.0, "reward": 0.8869513869285583, "reward_std": 0.583450973033905, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46139997243881226, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 310.1875, "completions/mean_terminated_length": 310.1875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.1884, "grad_norm": 0.09398400038480759, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10878376.0, "reward": 1.0160105228424072, "reward_std": 0.49912190437316895, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.532018780708313, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 573.0, "completions/mean_terminated_length": 573.0, "completions/min_length": 454.0, "completions/min_terminated_length": 454.0, "epoch": 0.1886, "grad_norm": 0.0712176188826561, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10892056.0, "reward": 0.962104082107544, "reward_std": 0.4892096519470215, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4867062568664551, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 254.1875, "completions/mean_terminated_length": 254.1875, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.1888, "grad_norm": 0.059048816561698914, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10899899.0, "reward": 0.5005073547363281, "reward_std": 0.13482333719730377, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06351249665021896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 438.3125, "completions/mean_terminated_length": 399.2666931152344, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.189, "grad_norm": 0.11689414083957672, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10910832.0, "reward": 0.9833573698997498, "reward_std": 0.5485090017318726, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.529212474822998, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9024999737739563, "step": 945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 363.875, "completions/mean_terminated_length": 363.875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1892, "grad_norm": 0.10150810331106186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10921310.0, "reward": 0.8983415365219116, "reward_std": 0.4805157482624054, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4216812551021576, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 730.4375, "completions/mean_terminated_length": 662.6923217773438, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.1894, "grad_norm": 0.07677993178367615, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10937077.0, "reward": 0.6173319816589355, "reward_std": 0.5528234243392944, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4846624732017517, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.43312501907348633, "step": 947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 372.125, "completions/mean_terminated_length": 372.125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1896, "grad_norm": 0.08172868937253952, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10947431.0, "reward": 0.6522667407989502, "reward_std": 0.3781917691230774, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2420312464237213, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8468749523162842, "step": 948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 470.5, "completions/mean_terminated_length": 470.5, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.1898, "grad_norm": 0.07600177824497223, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10959903.0, "reward": 0.7208322286605835, "reward_std": 0.38616836071014404, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2541624903678894, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 256.9375, "completions/mean_terminated_length": 256.9375, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.19, "grad_norm": 0.14441870152950287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10967822.0, "reward": 1.2568259239196777, "reward_std": 0.43392840027809143, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7636500000953674, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 670.0625, "completions/mean_terminated_length": 646.4666748046875, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.1902, "grad_norm": 0.01872815005481243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10982591.0, "reward": 0.533460259437561, "reward_std": 0.18430329859256744, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12941874563694, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 770.875, "completions/mean_terminated_length": 686.5, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.1904, "grad_norm": 0.056456249207258224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10998861.0, "reward": 0.547410249710083, "reward_std": 0.4967830777168274, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28231877088546753, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5768749713897705, "step": 952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 549.875, "completions/mean_terminated_length": 440.4615478515625, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.1906, "grad_norm": 0.04010825604200363, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11011659.0, "reward": 0.47390708327293396, "reward_std": 0.2610107362270355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1353124976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 278.125, "completions/mean_terminated_length": 278.125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.1908, "grad_norm": 0.11685030162334442, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11020109.0, "reward": 0.6793729066848755, "reward_std": 0.33112969994544983, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23374375700950623, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 450.9375, "completions/mean_terminated_length": 318.69232177734375, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.191, "grad_norm": 0.20045042037963867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11031516.0, "reward": 0.8245323896408081, "reward_std": 0.5955048203468323, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.46156251430511475, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.731249988079071, "step": 955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 866.0, "completions/mean_terminated_length": 855.4667358398438, "completions/min_length": 767.0, "completions/min_terminated_length": 767.0, "epoch": 0.1912, "grad_norm": 0.07004459947347641, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11049820.0, "reward": 0.8345977067947388, "reward_std": 0.5604020953178406, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41919374465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 325.0625, "completions/mean_terminated_length": 325.0625, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1914, "grad_norm": 0.22214624285697937, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11058765.0, "reward": 0.9526979923248291, "reward_std": 0.49854761362075806, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4678937494754791, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 342.125, "completions/mean_terminated_length": 342.125, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.1916, "grad_norm": 0.07785311341285706, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11068399.0, "reward": 1.377247929573059, "reward_std": 0.3330462872982025, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8794937133789062, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 492.5625, "completions/mean_terminated_length": 457.13336181640625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.1918, "grad_norm": 0.1072705015540123, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11080216.0, "reward": 0.8031132817268372, "reward_std": 0.5341365337371826, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41872498393058777, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8112499713897705, "step": 959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 446.375, "completions/mean_terminated_length": 446.375, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.192, "grad_norm": 0.003016711911186576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11093566.0, "reward": 0.5402822494506836, "reward_std": 0.015361327677965164, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08056250214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 246.625, "completions/mean_terminated_length": 246.625, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.1922, "grad_norm": 0.11177796870470047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11101384.0, "reward": 0.6602259874343872, "reward_std": 0.3276877999305725, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19545000791549683, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 647.375, "completions/mean_terminated_length": 622.2667236328125, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.1924, "grad_norm": 0.03353191539645195, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11115982.0, "reward": 0.6615631580352783, "reward_std": 0.24975967407226562, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3856250047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.65625, "step": 962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 893.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 576.5, "completions/mean_terminated_length": 576.5, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.1926, "grad_norm": 0.04997408762574196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11128742.0, "reward": 0.6324759721755981, "reward_std": 0.3649459779262543, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2024500072002411, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.953125, "step": 963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 291.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 227.0, "completions/mean_terminated_length": 227.0, "completions/min_length": 137.0, "completions/min_terminated_length": 137.0, "epoch": 0.1928, "grad_norm": 0.006640864070504904, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11136054.0, "reward": 0.5378011465072632, "reward_std": 0.0200816597789526, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07559999823570251, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 317.0, "completions/max_terminated_length": 317.0, "completions/mean_length": 246.75, "completions/mean_terminated_length": 246.75, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.193, "grad_norm": 0.1094534620642662, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11143762.0, "reward": 0.6614729762077332, "reward_std": 0.32745981216430664, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19794374704360962, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 506.75, "completions/mean_terminated_length": 506.75, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.1932, "grad_norm": 0.0374600924551487, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11158766.0, "reward": 0.5684009194374084, "reward_std": 0.2828383147716522, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13679999113082886, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 324.875, "completions/mean_terminated_length": 324.875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1934, "grad_norm": 0.06944067031145096, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11167596.0, "reward": 1.4398229122161865, "reward_std": 0.23761995136737823, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421437978744507, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 448.6875, "completions/mean_terminated_length": 448.6875, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.1936, "grad_norm": 0.07900723814964294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11178423.0, "reward": 1.1947135925292969, "reward_std": 0.46757081151008606, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7019249796867371, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 805.8125, "completions/mean_terminated_length": 755.4615478515625, "completions/min_length": 557.0, "completions/min_terminated_length": 557.0, "epoch": 0.1938, "grad_norm": 0.06064968928694725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11195700.0, "reward": 0.7501538991928101, "reward_std": 0.5623634457588196, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.500306248664856, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5993750095367432, "step": 969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 288.0, "completions/max_terminated_length": 288.0, "completions/mean_length": 242.9375, "completions/mean_terminated_length": 242.9375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.194, "grad_norm": 0.1202826201915741, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11203043.0, "reward": 1.3803073167800903, "reward_std": 0.3246068060398102, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8856124877929688, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 469.25, "completions/mean_terminated_length": 432.2666931152344, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1942, "grad_norm": 0.04627067223191261, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11217351.0, "reward": 0.7693754434585571, "reward_std": 0.2654021382331848, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6012499928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.48750001192092896, "step": 971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 719.875, "completions/mean_terminated_length": 618.5, "completions/min_length": 488.0, "completions/min_terminated_length": 488.0, "epoch": 0.1944, "grad_norm": 0.058435216546058655, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11232581.0, "reward": 0.43451958894729614, "reward_std": 0.38377296924591064, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11903749406337738, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7143750190734863, "step": 972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 209.0, "completions/max_terminated_length": 209.0, "completions/mean_length": 134.375, "completions/mean_terminated_length": 134.375, "completions/min_length": 100.0, "completions/min_terminated_length": 100.0, "epoch": 0.1946, "grad_norm": 0.09688491374254227, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11238811.0, "reward": 1.43802011013031, "reward_std": 0.24748528003692627, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9385374784469604, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9912500381469727, "step": 973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 480.6875, "completions/mean_terminated_length": 480.6875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1948, "grad_norm": 0.06489066779613495, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11250534.0, "reward": 0.6673448085784912, "reward_std": 0.32501113414764404, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2096875011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 268.125, "completions/mean_terminated_length": 268.125, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.195, "grad_norm": 0.06848226487636566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11258840.0, "reward": 1.4396040439605713, "reward_std": 0.23756134510040283, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9417062401771545, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 460.8125, "completions/mean_terminated_length": 460.8125, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.1952, "grad_norm": 0.07918371260166168, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11269989.0, "reward": 1.1110167503356934, "reward_std": 0.4670921564102173, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6595312356948853, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9031249284744263, "step": 976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 325.125, "completions/mean_terminated_length": 325.125, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.1954, "grad_norm": 0.1425054967403412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11279015.0, "reward": 1.0089733600616455, "reward_std": 0.5070760250091553, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5179437398910522, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 462.3125, "completions/mean_terminated_length": 462.3125, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.1956, "grad_norm": 0.06108576059341431, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11291436.0, "reward": 0.4336731731891632, "reward_std": 0.21617865562438965, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05484374612569809, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.890625, "step": 978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 704.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 488.4375, "completions/mean_terminated_length": 488.4375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.1958, "grad_norm": 0.05647144094109535, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11304003.0, "reward": 0.7244541645050049, "reward_std": 0.29259029030799866, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5114062428474426, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.518750011920929, "step": 979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 524.1875, "completions/mean_terminated_length": 524.1875, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.196, "grad_norm": 0.04457992687821388, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11316486.0, "reward": 0.6941413879394531, "reward_std": 0.2322217971086502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4507812559604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6281249523162842, "step": 980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.1962, "grad_norm": 0.09672840684652328, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11326619.0, "reward": 0.9050665497779846, "reward_std": 0.47508755326271057, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4351312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 483.625, "completions/mean_terminated_length": 483.625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.1964, "grad_norm": 0.07445422559976578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11339125.0, "reward": 1.2020981311798096, "reward_std": 0.4556112587451935, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7166937589645386, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 973.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 761.9375, "completions/mean_terminated_length": 761.9375, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.1966, "grad_norm": 0.02268151380121708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11355684.0, "reward": 0.7630443572998047, "reward_std": 0.18847328424453735, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5260875225067139, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6150000095367432, "step": 983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 432.4375, "completions/mean_terminated_length": 432.4375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.1968, "grad_norm": 0.09007880836725235, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11366443.0, "reward": 0.6332041025161743, "reward_std": 0.3644181787967682, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20390623807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 620.9375, "completions/mean_terminated_length": 594.0667114257812, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.197, "grad_norm": 0.04968668892979622, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11380586.0, "reward": 0.8117974996566772, "reward_std": 0.35475805401802063, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7485937476158142, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.171875, "step": 985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 263.0, "completions/max_terminated_length": 263.0, "completions/mean_length": 196.0625, "completions/mean_terminated_length": 196.0625, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.1972, "grad_norm": 0.07830638438463211, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11387627.0, "reward": 0.5887198448181152, "reward_std": 0.24347305297851562, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11493749916553497, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 605.125, "completions/mean_terminated_length": 577.2000122070312, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1974, "grad_norm": 0.09124529361724854, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11401693.0, "reward": 0.8844136595726013, "reward_std": 0.5847488641738892, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4563249945640564, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 799.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 594.6875, "completions/mean_terminated_length": 594.6875, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.1976, "grad_norm": 0.08745039254426956, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11414872.0, "reward": 0.8889670372009277, "reward_std": 0.4889916181564331, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.402931272983551, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750619888306, "step": 988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 295.5625, "completions/mean_terminated_length": 295.5625, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.1978, "grad_norm": 0.07316920161247253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11423073.0, "reward": 0.5653603076934814, "reward_std": 0.2833018898963928, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13071875274181366, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 352.9375, "completions/mean_terminated_length": 352.9375, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.198, "grad_norm": 0.0958554670214653, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11432496.0, "reward": 0.9049197435379028, "reward_std": 0.4755415916442871, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4348374903202057, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 832.0, "completions/mean_terminated_length": 804.5714721679688, "completions/min_length": 550.0, "completions/min_terminated_length": 550.0, "epoch": 0.1982, "grad_norm": 0.04158443212509155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11450064.0, "reward": 0.4886603057384491, "reward_std": 0.37719348073005676, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16481873393058777, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8112499713897705, "step": 991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 291.5625, "completions/mean_terminated_length": 291.5625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1984, "grad_norm": 0.09170696884393692, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11458537.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 511.3125, "completions/mean_terminated_length": 477.13336181640625, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1986, "grad_norm": 0.031092461198568344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11471534.0, "reward": 0.49078211188316345, "reward_std": 0.19619517028331757, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10656249523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.856249988079071, "step": 993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 342.25, "completions/mean_terminated_length": 342.25, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1988, "grad_norm": 0.08158724009990692, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11480946.0, "reward": 1.4397165775299072, "reward_std": 0.23759140074253082, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419312477111816, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 713.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 475.8125, "completions/mean_terminated_length": 475.8125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.199, "grad_norm": 0.0909993052482605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11492783.0, "reward": 0.8079694509506226, "reward_std": 0.41925016045570374, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6159375309944153, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.515625, "step": 995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 859.0625, "completions/mean_terminated_length": 784.0909423828125, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.1992, "grad_norm": 0.06565120071172714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11511088.0, "reward": 0.539910078048706, "reward_std": 0.5516161322593689, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2673187255859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6362500190734863, "step": 996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 296.0, "completions/max_terminated_length": 296.0, "completions/mean_length": 223.0625, "completions/mean_terminated_length": 223.0625, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.1994, "grad_norm": 0.1926446259021759, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11518625.0, "reward": 0.94356369972229, "reward_std": 0.5067483186721802, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4496249854564667, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.987500011920929, "step": 997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 306.0625, "completions/mean_terminated_length": 306.0625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1996, "grad_norm": 0.31868019700050354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11528082.0, "reward": 0.5819759964942932, "reward_std": 0.5100744962692261, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22644999623298645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 472.4375, "completions/mean_terminated_length": 435.66668701171875, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1998, "grad_norm": 0.08651988953351974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11539273.0, "reward": 1.1228915452957153, "reward_std": 0.48015129566192627, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6832812428474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 628.4375, "completions/mean_terminated_length": 628.4375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.2, "grad_norm": 0.06974131613969803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11553744.0, "reward": 0.9198229312896729, "reward_std": 0.543102502822876, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4646437466144562, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 1000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 374.9375, "completions/mean_terminated_length": 374.9375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.2002, "grad_norm": 0.005029667634516954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11563407.0, "reward": 0.5414697527885437, "reward_std": 0.016065986827015877, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08293749392032623, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 376.625, "completions/mean_terminated_length": 376.625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.2004, "grad_norm": 0.10565735399723053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11572745.0, "reward": 0.700626015663147, "reward_std": 0.4212134778499603, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2762500047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9375, "step": 1002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 443.75, "completions/mean_terminated_length": 443.75, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.2006, "grad_norm": 0.13981321454048157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11583893.0, "reward": 1.2269541025161743, "reward_std": 0.5014495253562927, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7664062976837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 281.1875, "completions/mean_terminated_length": 281.1875, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.2008, "grad_norm": 0.06480810046195984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11592536.0, "reward": 0.5856888890266418, "reward_std": 0.24490132927894592, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10887499153614044, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 1004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 639.5625, "completions/mean_terminated_length": 613.933349609375, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.201, "grad_norm": 0.03290551155805588, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11607425.0, "reward": 0.58961021900177, "reward_std": 0.2853762209415436, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17921873927116394, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.871874988079071, "step": 1005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 434.25, "completions/mean_terminated_length": 434.25, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.2012, "grad_norm": 0.08585038781166077, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11620693.0, "reward": 1.1657696962356567, "reward_std": 0.525234580039978, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7065374851226807, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 711.3125, "completions/mean_terminated_length": 666.6428833007812, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.2014, "grad_norm": 0.06962534785270691, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11636426.0, "reward": 0.7067042589187622, "reward_std": 0.5053190588951111, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28840625286102295, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8481249809265137, "step": 1007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 467.25, "completions/mean_terminated_length": 467.25, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.2016, "grad_norm": 0.04207586497068405, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11647374.0, "reward": 1.3977351188659668, "reward_std": 0.20636674761772156, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.8579687476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 1008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 284.8125, "completions/mean_terminated_length": 284.8125, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.2018, "grad_norm": 0.20363155007362366, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11655531.0, "reward": 0.7820290327072144, "reward_std": 0.38633114099502563, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31405624747276306, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8431249856948853, "step": 1009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 484.25, "completions/mean_terminated_length": 484.25, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.202, "grad_norm": 0.05835241451859474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11667391.0, "reward": 0.9452850222587585, "reward_std": 0.31959325075149536, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.703068733215332, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.596875011920929, "step": 1010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 301.0, "completions/max_terminated_length": 301.0, "completions/mean_length": 205.5, "completions/mean_terminated_length": 205.5, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 0.2022, "grad_norm": 0.13155750930309296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11674311.0, "reward": 1.3163011074066162, "reward_std": 0.39373189210891724, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8200997114181519, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9786250591278076, "step": 1011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 329.75, "completions/mean_terminated_length": 329.75, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.2024, "grad_norm": 0.0991869568824768, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11682963.0, "reward": 0.6566261649131775, "reward_std": 0.3295508027076721, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18825000524520874, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 204.5, "completions/mean_terminated_length": 204.5, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.2026, "grad_norm": 0.16783110797405243, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11690027.0, "reward": 1.3118669986724854, "reward_std": 0.4027816653251648, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.811231255531311, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731187224388123, "step": 1013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 326.75, "completions/mean_terminated_length": 326.75, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.2028, "grad_norm": 0.10217870026826859, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11699175.0, "reward": 0.665075957775116, "reward_std": 0.3255765438079834, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20515000820159912, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 341.5, "completions/mean_terminated_length": 341.5, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.203, "grad_norm": 0.15142202377319336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11709263.0, "reward": 1.1182541847229004, "reward_std": 0.52839195728302, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6740062236785889, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468750357627869, "step": 1015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 577.1875, "completions/mean_terminated_length": 547.4000244140625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.2032, "grad_norm": 0.05212901160120964, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11722914.0, "reward": 0.6300759315490723, "reward_std": 0.400535523891449, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2601500153541565, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8556249737739563, "step": 1016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 721.625, "completions/mean_terminated_length": 678.4285888671875, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.2034, "grad_norm": 0.0344083234667778, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11739036.0, "reward": 0.48889482021331787, "reward_std": 0.34278127551078796, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10278750211000443, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8506250381469727, "step": 1017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 292.375, "completions/mean_terminated_length": 292.375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.2036, "grad_norm": 0.17542125284671783, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11747586.0, "reward": 0.954513669013977, "reward_std": 0.4969290792942047, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47152501344680786, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 1018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 708.0, "completions/mean_length": 499.625, "completions/mean_terminated_length": 464.66668701171875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.2038, "grad_norm": 0.09188439697027206, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11759932.0, "reward": 0.6131258606910706, "reward_std": 0.43541809916496277, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2887499928474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 1019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 899.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 524.6875, "completions/mean_terminated_length": 524.6875, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.204, "grad_norm": 0.10450984537601471, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11772087.0, "reward": 0.7064695954322815, "reward_std": 0.4676588177680969, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4129374921321869, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6449999809265137, "step": 1020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 432.0625, "completions/mean_terminated_length": 432.0625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.2042, "grad_norm": 0.10635204613208771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11783192.0, "reward": 1.258591890335083, "reward_std": 0.4306727647781372, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7671812772750854, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 1021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 904.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 554.4375, "completions/mean_terminated_length": 554.4375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.2044, "grad_norm": 0.0904378741979599, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11796223.0, "reward": 1.1403415203094482, "reward_std": 0.4791785478591919, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6556812524795532, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 1022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 325.125, "completions/mean_terminated_length": 325.125, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.2046, "grad_norm": 0.12292419373989105, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11805617.0, "reward": 0.8947479128837585, "reward_std": 0.4836634397506714, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41449376940727234, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 982.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 604.25, "completions/mean_terminated_length": 604.25, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.2048, "grad_norm": 0.03082236275076866, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11819045.0, "reward": 0.6095322370529175, "reward_std": 0.1556246131658554, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21906250715255737, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 1024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 391.75, "completions/mean_terminated_length": 349.6000061035156, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.205, "grad_norm": 0.10614731907844543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11828977.0, "reward": 1.049447774887085, "reward_std": 0.5428796410560608, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5988937616348267, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 819.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 643.0625, "completions/mean_terminated_length": 643.0625, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.2052, "grad_norm": 0.07672005891799927, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11843074.0, "reward": 1.1363823413848877, "reward_std": 0.5781172513961792, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7102625370025635, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 1026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 763.0, "completions/mean_length": 550.5, "completions/mean_terminated_length": 518.933349609375, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.2054, "grad_norm": 0.02038111910223961, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11855882.0, "reward": 0.5299228429794312, "reward_std": 0.15241850912570953, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1223437488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8718749284744263, "step": 1027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 277.0, "completions/max_terminated_length": 277.0, "completions/mean_length": 174.25, "completions/mean_terminated_length": 174.25, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "epoch": 0.2056, "grad_norm": 0.2727687656879425, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11862686.0, "reward": 0.7656793594360352, "reward_std": 0.4376787543296814, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28135624527931213, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 1028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 757.0, "completions/mean_terminated_length": 739.2000122070312, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.2058, "grad_norm": 0.05497273430228233, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11878702.0, "reward": 0.7786603569984436, "reward_std": 0.45911434292793274, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3698187470436096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8581249713897705, "step": 1029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 254.0, "completions/mean_terminated_length": 254.0, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.206, "grad_norm": 0.10693903267383575, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11886446.0, "reward": 0.7093479633331299, "reward_std": 0.39232057332992554, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23119375109672546, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 1030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 356.625, "completions/mean_terminated_length": 356.625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.2062, "grad_norm": 0.06431489437818527, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11896600.0, "reward": 1.3786641359329224, "reward_std": 0.33161357045173645, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8823249936103821, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9950000047683716, "step": 1031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 811.8125, "completions/mean_terminated_length": 684.5, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.2064, "grad_norm": 0.0651201382279396, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11913845.0, "reward": 0.5340505838394165, "reward_std": 0.5510803461074829, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2555999755859375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5800000429153442, "step": 1032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 706.25, "completions/mean_terminated_length": 660.857177734375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2066, "grad_norm": 0.10596515983343124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11929177.0, "reward": 1.0810039043426514, "reward_std": 0.5818655490875244, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6620062589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8143749833106995, "step": 1033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 744.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 401.875, "completions/mean_terminated_length": 401.875, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.2068, "grad_norm": 0.10584503412246704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11940375.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 302.1875, "completions/mean_terminated_length": 302.1875, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.207, "grad_norm": 0.36936748027801514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11948922.0, "reward": 0.668204665184021, "reward_std": 0.431997686624527, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21140626072883606, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 1035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 851.0, "completions/mean_length": 541.5625, "completions/mean_terminated_length": 509.4000244140625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2072, "grad_norm": 0.053315743803977966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11962035.0, "reward": 0.5602226853370667, "reward_std": 0.2694227993488312, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24544374644756317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7456250190734863, "step": 1036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 672.625, "completions/mean_terminated_length": 649.2000122070312, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.2074, "grad_norm": 0.05647699162364006, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11976749.0, "reward": 0.6375791430473328, "reward_std": 0.4014885723590851, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2751562297344208, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 1037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 479.0625, "completions/mean_terminated_length": 479.0625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2076, "grad_norm": 0.10791502147912979, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11988478.0, "reward": 0.9532010555267334, "reward_std": 0.49765798449516296, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46890002489089966, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 1038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 752.8125, "completions/mean_terminated_length": 752.8125, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.2078, "grad_norm": 0.028211895376443863, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12005115.0, "reward": 0.5851199626922607, "reward_std": 0.24485595524311066, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10773749649524689, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 1039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 508.5625, "completions/mean_terminated_length": 474.20001220703125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.208, "grad_norm": 0.03562250733375549, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12017012.0, "reward": 0.5337510108947754, "reward_std": 0.1765066534280777, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12999999523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.856249988079071, "step": 1040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 417.625, "completions/mean_terminated_length": 417.625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.2082, "grad_norm": 0.10185800492763519, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12027774.0, "reward": 0.7159759998321533, "reward_std": 0.38871315121650696, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24445000290870667, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 347.1875, "completions/mean_terminated_length": 347.1875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.2084, "grad_norm": 0.1188458502292633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12037121.0, "reward": 1.0173166990280151, "reward_std": 0.4978773593902588, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5346312522888184, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 1042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 227.0, "completions/max_terminated_length": 227.0, "completions/mean_length": 192.4375, "completions/mean_terminated_length": 192.4375, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.2086, "grad_norm": 0.00020654682884924114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12044360.0, "reward": 1.4990479946136475, "reward_std": 0.0005329799023456872, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980937242507935, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 235.6875, "completions/mean_terminated_length": 235.6875, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.2088, "grad_norm": 0.16005265712738037, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12051859.0, "reward": 1.140225887298584, "reward_std": 0.47844094038009644, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6554499864578247, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 1044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 399.0, "completions/max_terminated_length": 399.0, "completions/mean_length": 296.5, "completions/mean_terminated_length": 296.5, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.209, "grad_norm": 0.14109547436237335, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12060091.0, "reward": 0.8293575048446655, "reward_std": 0.46684530377388, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3462125062942505, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 1045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 281.0, "completions/max_terminated_length": 281.0, "completions/mean_length": 244.4375, "completions/mean_terminated_length": 244.4375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.2092, "grad_norm": 0.11467792093753815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12067634.0, "reward": 0.5821324586868286, "reward_std": 0.24540428817272186, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10176250338554382, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 1046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 646.1875, "completions/mean_terminated_length": 592.2142944335938, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.2094, "grad_norm": 0.043101802468299866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12083493.0, "reward": 0.5424226522445679, "reward_std": 0.3168862462043762, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14734375476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.828125, "step": 1047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 550.125, "completions/mean_terminated_length": 550.125, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.2096, "grad_norm": 0.08244792371988297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12096519.0, "reward": 0.9058446288108826, "reward_std": 0.47830313444137573, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5616875290870667, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7200000286102295, "step": 1048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 697.5625, "completions/mean_terminated_length": 675.800048828125, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.2098, "grad_norm": 0.03472369536757469, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12111888.0, "reward": 0.568972647190094, "reward_std": 0.24712242186069489, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26294374465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 1049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 374.125, "completions/mean_terminated_length": 374.125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.21, "grad_norm": 0.14250551164150238, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12121426.0, "reward": 1.1515166759490967, "reward_std": 0.46673280000686646, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6780312061309814, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 1050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 452.0, "completions/max_terminated_length": 452.0, "completions/mean_length": 282.6875, "completions/mean_terminated_length": 282.6875, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "epoch": 0.2102, "grad_norm": 0.1513727307319641, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12129981.0, "reward": 0.8739850521087646, "reward_std": 0.2744295299053192, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8104687929153442, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.17812499403953552, "step": 1051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 425.125, "completions/mean_terminated_length": 425.125, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.2104, "grad_norm": 0.020728008821606636, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12140863.0, "reward": 0.5739072561264038, "reward_std": 0.114273801445961, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14781248569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 1052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 752.375, "completions/mean_terminated_length": 734.2667236328125, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.2106, "grad_norm": 0.015019048005342484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12157285.0, "reward": 0.5109822750091553, "reward_std": 0.14612515270709991, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0844624936580658, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 1053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 495.25, "completions/mean_terminated_length": 460.0000305175781, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.2108, "grad_norm": 0.10155714303255081, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12169097.0, "reward": 0.8600603938102722, "reward_std": 0.5280404090881348, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40761876106262207, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9118750095367432, "step": 1054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 662.3125, "completions/mean_terminated_length": 610.6428833007812, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.211, "grad_norm": 0.06273991614580154, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12184206.0, "reward": 0.586094856262207, "reward_std": 0.3472119867801666, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23468750715255737, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7625000476837158, "step": 1055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 730.25, "completions/mean_terminated_length": 710.6666870117188, "completions/min_length": 562.0, "completions/min_terminated_length": 562.0, "epoch": 0.2112, "grad_norm": 0.05187653377652168, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12199986.0, "reward": 0.7691761255264282, "reward_std": 0.4377497136592865, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.413349986076355, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7262500524520874, "step": 1056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 372.5, "completions/mean_terminated_length": 372.5, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2114, "grad_norm": 0.13777174055576324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12210474.0, "reward": 0.9477358460426331, "reward_std": 0.4670316278934479, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5204687118530273, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 1057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 644.5, "completions/mean_terminated_length": 644.5, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.2116, "grad_norm": 0.03296738862991333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12224402.0, "reward": 0.509597897529602, "reward_std": 0.13658054172992706, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08169374614953995, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 520.5, "completions/mean_terminated_length": 520.5, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.2118, "grad_norm": 0.08447601646184921, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12236762.0, "reward": 1.0577353239059448, "reward_std": 0.38039684295654297, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7404687404632568, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6218750476837158, "step": 1059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 346.5625, "completions/mean_terminated_length": 346.5625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.212, "grad_norm": 0.00018616358283907175, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12246275.0, "reward": 1.4993979930877686, "reward_std": 0.0005986018804833293, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987937808036804, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 1060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 517.3125, "completions/mean_terminated_length": 517.3125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.2122, "grad_norm": 0.03190421685576439, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12259016.0, "reward": 0.6350666880607605, "reward_std": 0.18280088901519775, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.27013126015663147, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7893750071525574, "step": 1061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 449.4375, "completions/mean_terminated_length": 449.4375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.2124, "grad_norm": 0.08024519681930542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12270191.0, "reward": 0.82203209400177, "reward_std": 0.3618515729904175, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.45656248927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.793749988079071, "step": 1062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 308.0, "completions/max_terminated_length": 308.0, "completions/mean_length": 273.875, "completions/mean_terminated_length": 273.875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2126, "grad_norm": 0.18383930623531342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12279085.0, "reward": 0.8273949027061462, "reward_std": 0.5646774172782898, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40478748083114624, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 1063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 507.125, "completions/mean_terminated_length": 507.125, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.2128, "grad_norm": 0.05747630447149277, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12292415.0, "reward": 0.6675009727478027, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20999999344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 461.0625, "completions/mean_terminated_length": 461.0625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.213, "grad_norm": 0.07796580344438553, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12303568.0, "reward": 0.7238947749137878, "reward_std": 0.3848000764846802, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26028749346733093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 382.375, "completions/mean_terminated_length": 382.375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.2132, "grad_norm": 0.08275946229696274, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12313542.0, "reward": 1.3214166164398193, "reward_std": 0.3273216485977173, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8928312659263611, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 1066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 322.0, "completions/max_terminated_length": 322.0, "completions/mean_length": 260.375, "completions/mean_terminated_length": 260.375, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.2134, "grad_norm": 0.1586335301399231, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12321452.0, "reward": 1.1373918056488037, "reward_std": 0.48276591300964355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6497812271118164, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 1067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 936.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 695.75, "completions/mean_terminated_length": 695.75, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.2136, "grad_norm": 0.019156139343976974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12337016.0, "reward": 0.8290630578994751, "reward_std": 0.12075909227132797, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6581250429153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5374999642372131, "step": 1068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 224.0, "completions/mean_terminated_length": 224.0, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.2138, "grad_norm": 0.00018808216555044055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12344920.0, "reward": 1.4991042613983154, "reward_std": 0.0005441161920316517, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982062578201294, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 1069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 406.8125, "completions/mean_terminated_length": 406.8125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.214, "grad_norm": 0.1167440414428711, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12356261.0, "reward": 1.1675010919570923, "reward_std": 0.4550197124481201, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7099999785423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 1070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 176.0, "completions/max_terminated_length": 176.0, "completions/mean_length": 129.9375, "completions/mean_terminated_length": 129.9375, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "epoch": 0.2142, "grad_norm": 0.00012713621254079044, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12362212.0, "reward": 1.4998918771743774, "reward_std": 0.00030518282437697053, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9997812509536743, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9906250238418579, "step": 1071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 377.1875, "completions/mean_terminated_length": 334.0666809082031, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.2144, "grad_norm": 0.026724467054009438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12371719.0, "reward": 0.4938042163848877, "reward_std": 0.1333692967891693, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05010624974966049, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9118750095367432, "step": 1072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 456.3125, "completions/mean_terminated_length": 456.3125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.2146, "grad_norm": 0.0855763703584671, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12382556.0, "reward": 1.0138354301452637, "reward_std": 0.5017363429069519, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5276687145233154, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 1073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 453.5625, "completions/mean_terminated_length": 372.0714416503906, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2148, "grad_norm": 0.11792073398828506, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12393685.0, "reward": 0.6032041311264038, "reward_std": 0.4472525119781494, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.26890623569488525, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7625000476837158, "step": 1074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 360.875, "completions/mean_terminated_length": 360.875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.215, "grad_norm": 0.22208696603775024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12404195.0, "reward": 0.5989072322845459, "reward_std": 0.39724069833755493, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1978124976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 264.5625, "completions/mean_terminated_length": 264.5625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.2152, "grad_norm": 0.07721437513828278, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12412412.0, "reward": 1.4372353553771973, "reward_std": 0.24727636575698853, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9369688034057617, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 1076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 461.1875, "completions/mean_terminated_length": 461.1875, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.2154, "grad_norm": 0.09166645258665085, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12424767.0, "reward": 1.0784887075424194, "reward_std": 0.493402361869812, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5944750308990479, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 1077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 533.0, "completions/mean_terminated_length": 533.0, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.2156, "grad_norm": 0.0935039296746254, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12438239.0, "reward": 0.837807297706604, "reward_std": 0.4606812596321106, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36311250925064087, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 688.75, "completions/mean_terminated_length": 666.4000244140625, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.2158, "grad_norm": 0.05558918043971062, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12453179.0, "reward": 0.6677353382110596, "reward_std": 0.370360791683197, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27296873927116394, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 1079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 772.75, "completions/mean_terminated_length": 736.857177734375, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.216, "grad_norm": 0.07128676027059555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12469415.0, "reward": 0.925513505935669, "reward_std": 0.5752538442611694, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.4760249853134155, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8362500071525574, "step": 1080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 399.375, "completions/mean_terminated_length": 399.375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.2162, "grad_norm": 0.11586380749940872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12479821.0, "reward": 0.9051447510719299, "reward_std": 0.4751918613910675, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4352875053882599, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 982.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 743.0625, "completions/mean_terminated_length": 743.0625, "completions/min_length": 568.0, "completions/min_terminated_length": 568.0, "epoch": 0.2164, "grad_norm": 0.0486731082201004, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12495902.0, "reward": 0.7786601185798645, "reward_std": 0.2642419934272766, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4948187470436096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6643749475479126, "step": 1082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 303.4375, "completions/mean_terminated_length": 303.4375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2166, "grad_norm": 0.0976376011967659, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12504613.0, "reward": 1.3752950429916382, "reward_std": 0.33974283933639526, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8755874633789062, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 1083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 330.9375, "completions/mean_terminated_length": 330.9375, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.2168, "grad_norm": 0.10873962193727493, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12513268.0, "reward": 0.6774229407310486, "reward_std": 0.3253090977668762, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22984373569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 1084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 585.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 398.0625, "completions/mean_terminated_length": 398.0625, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.217, "grad_norm": 0.10128449648618698, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12524069.0, "reward": 0.9188885688781738, "reward_std": 0.4717354476451874, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.46277502179145813, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9350000023841858, "step": 1085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 599.0625, "completions/mean_terminated_length": 599.0625, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.2172, "grad_norm": 0.07485698163509369, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12537366.0, "reward": 0.7028133869171143, "reward_std": 0.36772421002388, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.34312498569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.800000011920929, "step": 1086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 521.5625, "completions/mean_terminated_length": 521.5625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.2174, "grad_norm": 0.032571710646152496, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12550751.0, "reward": 1.0539847612380981, "reward_std": 0.2286052256822586, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.9204687476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3656250238418579, "step": 1087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 320.0, "completions/mean_terminated_length": 320.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2176, "grad_norm": 0.10587809979915619, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12559471.0, "reward": 0.7844415903091431, "reward_std": 0.4259919226169586, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31888124346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 384.0625, "completions/mean_terminated_length": 384.0625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.2178, "grad_norm": 0.2293284833431244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12569888.0, "reward": 0.9265353679656982, "reward_std": 0.5378381609916687, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4780687689781189, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 318.9375, "completions/mean_terminated_length": 318.9375, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.218, "grad_norm": 0.005014030262827873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12579615.0, "reward": 0.5485947132110596, "reward_std": 0.017349589616060257, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 461.3125, "completions/mean_terminated_length": 461.3125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2182, "grad_norm": 0.06727702170610428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12590740.0, "reward": 0.7897664904594421, "reward_std": 0.2214498519897461, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5795312523841858, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.49687501788139343, "step": 1091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 322.0, "completions/max_terminated_length": 322.0, "completions/mean_length": 270.9375, "completions/mean_terminated_length": 270.9375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.2184, "grad_norm": 0.01274214219301939, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12599379.0, "reward": 0.5551415681838989, "reward_std": 0.045158982276916504, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11028124392032623, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 1092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 400.125, "completions/mean_terminated_length": 400.125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.2186, "grad_norm": 0.10601583868265152, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12609333.0, "reward": 1.2743604183197021, "reward_std": 0.40456151962280273, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7987188100814819, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 1093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 305.875, "completions/mean_terminated_length": 258.0, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.2188, "grad_norm": 0.019016509875655174, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12618291.0, "reward": 0.49199485778808594, "reward_std": 0.13276919722557068, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04648749530315399, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9137499928474426, "step": 1094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 352.6875, "completions/mean_terminated_length": 352.6875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.219, "grad_norm": 0.11919169872999191, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12627806.0, "reward": 0.7027698159217834, "reward_std": 0.39562374353408813, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21803750097751617, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 1095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 350.25, "completions/mean_terminated_length": 350.25, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.2192, "grad_norm": 0.07985420525074005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12637138.0, "reward": 1.2401355504989624, "reward_std": 0.3467560410499573, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8552687168121338, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6968750357627869, "step": 1096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 540.5, "completions/mean_terminated_length": 540.5, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2194, "grad_norm": 0.060754548758268356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12651050.0, "reward": 0.6788289546966553, "reward_std": 0.3120670020580292, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35765624046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7406250238418579, "step": 1097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 418.25, "completions/mean_terminated_length": 418.25, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.2196, "grad_norm": 0.06566612422466278, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12661294.0, "reward": 1.380054235458374, "reward_std": 0.32450953125953674, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851062059402466, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 328.0, "completions/max_terminated_length": 328.0, "completions/mean_length": 266.8125, "completions/mean_terminated_length": 266.8125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.2198, "grad_norm": 0.16131800413131714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12669067.0, "reward": 1.2595947980880737, "reward_std": 0.42823007702827454, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691874504089355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 345.0, "completions/max_terminated_length": 345.0, "completions/mean_length": 253.5, "completions/mean_terminated_length": 253.5, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.22, "grad_norm": 0.08610319346189499, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12676803.0, "reward": 0.6512540578842163, "reward_std": 0.33144912123680115, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1775062531232834, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 414.4375, "completions/mean_terminated_length": 414.4375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.2202, "grad_norm": 0.045684970915317535, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12687290.0, "reward": 0.6045072078704834, "reward_std": 0.23866796493530273, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1465124934911728, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 966.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 641.0625, "completions/mean_terminated_length": 641.0625, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.2204, "grad_norm": 0.07741378992795944, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12702379.0, "reward": 0.6807039380073547, "reward_std": 0.37284061312675476, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.42390623688697815, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6437499523162842, "step": 1102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 557.4375, "completions/mean_terminated_length": 526.3333740234375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.2206, "grad_norm": 0.05861556529998779, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12714866.0, "reward": 0.5212512016296387, "reward_std": 0.36577892303466797, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16749998927116394, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 1103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 289.5625, "completions/mean_terminated_length": 289.5625, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.2208, "grad_norm": 0.15149125456809998, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12723691.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 522.875, "completions/mean_terminated_length": 522.875, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.221, "grad_norm": 0.0207708477973938, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12736889.0, "reward": 0.589844822883606, "reward_std": 0.11623655259609222, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1796875, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 1105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 340.0625, "completions/mean_terminated_length": 340.0625, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.2212, "grad_norm": 0.10982061922550201, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12746538.0, "reward": 1.3462855815887451, "reward_std": 0.4302581548690796, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8800687193870544, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 666.1875, "completions/mean_terminated_length": 666.1875, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.2214, "grad_norm": 0.060051366686820984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12761485.0, "reward": 0.7292167544364929, "reward_std": 0.33854570984840393, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3334312438964844, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8368750214576721, "step": 1107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 584.75, "completions/mean_terminated_length": 584.75, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.2216, "grad_norm": 0.061376504600048065, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12775369.0, "reward": 0.5588165521621704, "reward_std": 0.3228761851787567, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18013125658035278, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9268749952316284, "step": 1108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 442.0, "completions/max_terminated_length": 442.0, "completions/mean_length": 331.4375, "completions/mean_terminated_length": 331.4375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.2218, "grad_norm": 0.14044256508350372, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12784288.0, "reward": 0.877225935459137, "reward_std": 0.5232343673706055, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4419499933719635, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 1109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 614.0, "completions/mean_terminated_length": 586.6666870117188, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.222, "grad_norm": 0.07066784054040909, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12799216.0, "reward": 0.6672697067260742, "reward_std": 0.3746100664138794, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2720375061035156, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8299999237060547, "step": 1110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 388.9375, "completions/mean_terminated_length": 388.9375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.2222, "grad_norm": 0.10940419882535934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12809375.0, "reward": 0.7710951566696167, "reward_std": 0.4346086084842682, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29218751192092896, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 1111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 265.0, "completions/max_terminated_length": 265.0, "completions/mean_length": 213.6875, "completions/mean_terminated_length": 213.6875, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.2224, "grad_norm": 0.4563845992088318, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12816442.0, "reward": 1.2232446670532227, "reward_std": 0.5079613924026489, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7589874863624573, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.971875011920929, "step": 1112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 305.8125, "completions/mean_terminated_length": 305.8125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.2226, "grad_norm": 0.12888377904891968, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12825351.0, "reward": 1.1312294006347656, "reward_std": 0.4908547103404999, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6374562382698059, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9718749523162842, "step": 1113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 283.4375, "completions/mean_terminated_length": 283.4375, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.2228, "grad_norm": 0.17081309854984283, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12833966.0, "reward": 1.1530447006225586, "reward_std": 0.4639458954334259, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6810874938964844, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "step": 1114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 493.8125, "completions/mean_terminated_length": 493.8125, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.223, "grad_norm": 0.1272912472486496, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12846075.0, "reward": 0.9271105527877808, "reward_std": 0.538030743598938, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4792187213897705, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9031250476837158, "step": 1115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 322.8125, "completions/mean_terminated_length": 322.8125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.2232, "grad_norm": 0.1868409663438797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12855256.0, "reward": 0.8833606243133545, "reward_std": 0.5250858664512634, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4542187452316284, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 1116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 254.625, "completions/mean_terminated_length": 254.625, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.2234, "grad_norm": 0.037402912974357605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12862690.0, "reward": 1.4739699363708496, "reward_std": 0.10106033086776733, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9479374885559082, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9137499928474426, "step": 1117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 483.6875, "completions/mean_terminated_length": 447.66668701171875, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.2236, "grad_norm": 0.10281744599342346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12874397.0, "reward": 1.015000820159912, "reward_std": 0.5271903276443481, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5924999713897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.856249988079071, "step": 1118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 679.1875, "completions/mean_terminated_length": 656.2000122070312, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.2238, "grad_norm": 0.042744819074869156, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12889456.0, "reward": 0.5890634059906006, "reward_std": 0.22048789262771606, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24062499403953552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 1119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 356.625, "completions/mean_terminated_length": 356.625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.224, "grad_norm": 0.10322766751050949, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12899290.0, "reward": 0.5989072322845459, "reward_std": 0.39724069833755493, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1978124976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 333.75, "completions/mean_terminated_length": 333.75, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.2242, "grad_norm": 0.13629719614982605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12908470.0, "reward": 0.8430447578430176, "reward_std": 0.45700326561927795, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3735874891281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 321.1875, "completions/mean_terminated_length": 321.1875, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.2244, "grad_norm": 0.13094168901443481, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12917337.0, "reward": 1.0831791162490845, "reward_std": 0.48741284012794495, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6038562655448914, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 1122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 445.75, "completions/mean_terminated_length": 445.75, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.2246, "grad_norm": 0.0579308457672596, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12928389.0, "reward": 0.6729664206504822, "reward_std": 0.26647812128067017, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4084312319755554, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5993750095367432, "step": 1123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 841.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 622.5, "completions/mean_terminated_length": 622.5, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.2248, "grad_norm": 0.07784030586481094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12942429.0, "reward": 1.287360429763794, "reward_std": 0.393765926361084, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.824718713760376, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 1124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 377.9375, "completions/mean_terminated_length": 377.9375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.225, "grad_norm": 0.0471811480820179, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12952300.0, "reward": 0.6232665777206421, "reward_std": 0.18040145933628082, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24653124809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7981250286102295, "step": 1125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 526.5625, "completions/mean_terminated_length": 526.5625, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.2252, "grad_norm": 0.03129686787724495, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12964821.0, "reward": 0.6507819890975952, "reward_std": 0.1449364274740219, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.30156248807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 1126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 392.5, "completions/mean_terminated_length": 392.5, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.2254, "grad_norm": 0.05723050981760025, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12975021.0, "reward": 0.7757040858268738, "reward_std": 0.2756304144859314, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.48890623450279236, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6468750238418579, "step": 1127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 412.9375, "completions/mean_terminated_length": 412.9375, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.2256, "grad_norm": 0.00016805219638627023, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12986332.0, "reward": 1.4992949962615967, "reward_std": 0.0006373135838657618, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985874891281128, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 1128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 408.6875, "completions/mean_terminated_length": 408.6875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.2258, "grad_norm": 0.09206736087799072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12996903.0, "reward": 1.326413631439209, "reward_std": 0.3736988306045532, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8403245210647583, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9671937227249146, "step": 1129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 539.75, "completions/mean_terminated_length": 539.75, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.226, "grad_norm": 0.09024295210838318, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13009603.0, "reward": 0.7950633764266968, "reward_std": 0.5325395464897156, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4026249945163727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8668749332427979, "step": 1130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 406.1875, "completions/mean_terminated_length": 406.1875, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.2262, "grad_norm": 0.09903870522975922, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13020246.0, "reward": 0.8936229944229126, "reward_std": 0.4846753776073456, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4122437536716461, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 367.875, "completions/mean_terminated_length": 367.875, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.2264, "grad_norm": 0.07962296903133392, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13030820.0, "reward": 0.65975421667099, "reward_std": 0.32828548550605774, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19450624287128448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 1132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 385.0, "completions/mean_terminated_length": 385.0, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.2266, "grad_norm": 0.11291240900754929, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13040420.0, "reward": 1.192063570022583, "reward_std": 0.47104644775390625, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6966249942779541, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 1133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 424.375, "completions/mean_terminated_length": 424.375, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.2268, "grad_norm": 0.047886621206998825, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13051978.0, "reward": 0.5738290548324585, "reward_std": 0.2820831835269928, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 306.0, "completions/max_terminated_length": 306.0, "completions/mean_length": 225.0, "completions/mean_terminated_length": 225.0, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.227, "grad_norm": 0.0002410965389572084, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13059306.0, "reward": 1.4993916749954224, "reward_std": 0.0005869506276212633, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987812638282776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 1135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 882.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 445.0, "completions/mean_terminated_length": 445.0, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.2272, "grad_norm": 0.1058000922203064, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13070810.0, "reward": 0.9339727759361267, "reward_std": 0.46495744585990906, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49294376373291016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 315.25, "completions/mean_terminated_length": 315.25, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2274, "grad_norm": 0.1551671028137207, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13079406.0, "reward": 0.8482823371887207, "reward_std": 0.4538612365722656, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3840624690055847, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 495.0625, "completions/mean_terminated_length": 495.0625, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.2276, "grad_norm": 0.08365224301815033, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13091311.0, "reward": 0.7498447299003601, "reward_std": 0.38099831342697144, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3121874928474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 1138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 291.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 180.75, "completions/mean_terminated_length": 180.75, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.2278, "grad_norm": 0.06854464113712311, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13097659.0, "reward": 0.9717978239059448, "reward_std": 0.11281251162290573, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9435937404632568, "rewards/confidence_one_or_zero": 0.9375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.05937499925494194, "step": 1139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 874.4375, "completions/mean_terminated_length": 758.1111450195312, "completions/min_length": 539.0, "completions/min_terminated_length": 539.0, "epoch": 0.228, "grad_norm": 0.11862742900848389, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13116002.0, "reward": 0.7816411852836609, "reward_std": 0.7482982873916626, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.500781238079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5218750238418579, "step": 1140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 510.0625, "completions/mean_terminated_length": 510.0625, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.2282, "grad_norm": 0.02620586007833481, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13127779.0, "reward": 0.6188291907310486, "reward_std": 0.1252298206090927, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23765625059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 1141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 513.625, "completions/mean_terminated_length": 513.625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.2284, "grad_norm": 0.018596196547150612, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13141725.0, "reward": 0.5882821083068848, "reward_std": 0.09318415075540543, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17656248807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 857.6875, "completions/mean_terminated_length": 728.3333129882812, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.2286, "grad_norm": 0.051373668015003204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13159400.0, "reward": 0.3197630047798157, "reward_std": 0.29455843567848206, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07702499628067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5212500095367432, "step": 1143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 963.9375, "completions/mean_terminated_length": 903.875, "completions/min_length": 820.0, "completions/min_terminated_length": 820.0, "epoch": 0.2288, "grad_norm": 0.03539091348648071, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13179847.0, "reward": 0.2725660800933838, "reward_std": 0.28159186244010925, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.045131247490644455, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4768750071525574, "step": 1144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 417.4375, "completions/mean_terminated_length": 377.0000305175781, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.229, "grad_norm": 0.18603551387786865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13190270.0, "reward": 1.025704026222229, "reward_std": 0.6681415438652039, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6764062643051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 1145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 606.375, "completions/mean_terminated_length": 606.375, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.2292, "grad_norm": 0.09464626014232635, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13203780.0, "reward": 1.0293759107589722, "reward_std": 0.5029418468475342, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6212500333786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 1146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 250.25, "completions/mean_terminated_length": 250.25, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.2294, "grad_norm": 0.19484815001487732, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13211656.0, "reward": 1.1092636585235596, "reward_std": 0.536721408367157, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6560249924659729, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 1147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 430.9375, "completions/mean_terminated_length": 430.9375, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.2296, "grad_norm": 0.09106837213039398, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13222295.0, "reward": 1.2644259929656982, "reward_std": 0.41997918486595154, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7788499593734741, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 482.625, "completions/mean_terminated_length": 482.625, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.2298, "grad_norm": 0.02680203504860401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13234593.0, "reward": 0.5083605647087097, "reward_std": 0.13657139241695404, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 1149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 535.4375, "completions/mean_terminated_length": 535.4375, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.23, "grad_norm": 0.04970502108335495, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13247528.0, "reward": 1.4380794763565063, "reward_std": 0.2475009709596634, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.938656210899353, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9931249618530273, "step": 1150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 441.5, "completions/mean_terminated_length": 441.5, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.2302, "grad_norm": 0.08644662797451019, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13258432.0, "reward": 0.7626572251319885, "reward_std": 0.3838670253753662, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.33781248331069946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 1151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 758.1875, "completions/mean_terminated_length": 740.4666748046875, "completions/min_length": 527.0, "completions/min_terminated_length": 527.0, "epoch": 0.2304, "grad_norm": 0.04483189061284065, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13274819.0, "reward": 0.7310167551040649, "reward_std": 0.39365479350090027, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3995312452316284, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.715624988079071, "step": 1152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 378.75, "completions/mean_terminated_length": 378.75, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.2306, "grad_norm": 0.007539704442024231, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13284687.0, "reward": 0.5317137241363525, "reward_std": 0.02290913462638855, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06342499703168869, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 378.1875, "completions/mean_terminated_length": 378.1875, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.2308, "grad_norm": 0.07681600749492645, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13294674.0, "reward": 0.5561700463294983, "reward_std": 0.2847377061843872, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11233749985694885, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9087499976158142, "step": 1154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 523.0625, "completions/mean_terminated_length": 523.0625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.231, "grad_norm": 0.1031150072813034, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13307091.0, "reward": 0.930210292339325, "reward_std": 0.5350662469863892, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48541873693466187, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943750262260437, "step": 1155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 450.25, "completions/mean_terminated_length": 450.25, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.2312, "grad_norm": 0.06431920826435089, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13317767.0, "reward": 0.6708447337150574, "reward_std": 0.3266097903251648, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21668750047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9450000524520874, "step": 1156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 676.0, "completions/mean_length": 501.25, "completions/mean_terminated_length": 466.4000244140625, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.2314, "grad_norm": 0.04087318107485771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13330123.0, "reward": 0.48015713691711426, "reward_std": 0.18743397295475006, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08531250059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.890625, "step": 1157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 612.5, "completions/mean_terminated_length": 553.7142944335938, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.2316, "grad_norm": 0.06329245865345001, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13344403.0, "reward": 0.5485948324203491, "reward_std": 0.32373449206352234, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15968748927116394, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 1158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 418.5, "completions/mean_terminated_length": 418.5, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.2318, "grad_norm": 0.08842403441667557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13355323.0, "reward": 0.7845821380615234, "reward_std": 0.42624354362487793, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3191624879837036, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 506.625, "completions/mean_terminated_length": 506.625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.232, "grad_norm": 0.10416356474161148, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13368597.0, "reward": 1.2593634128570557, "reward_std": 0.38596346974372864, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8312250375747681, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8487499952316284, "step": 1160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 613.9375, "completions/mean_terminated_length": 613.9375, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.2322, "grad_norm": 0.04243236035108566, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13382260.0, "reward": 0.6276572346687317, "reward_std": 0.24316434562206268, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1928125023841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 1161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 348.625, "completions/mean_terminated_length": 348.625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.2324, "grad_norm": 0.059463851153850555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13391822.0, "reward": 0.601460337638855, "reward_std": 0.23973597586154938, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14041873812675476, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 946.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 591.1875, "completions/mean_terminated_length": 591.1875, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.2326, "grad_norm": 0.07129152864217758, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13405169.0, "reward": 0.6876571178436279, "reward_std": 0.23345167934894562, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3128125071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.831250011920929, "step": 1163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 284.9375, "completions/mean_terminated_length": 284.9375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.2328, "grad_norm": 0.1535685956478119, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13414032.0, "reward": 0.9615572690963745, "reward_std": 0.48991647362709045, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48561251163482666, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 403.875, "completions/mean_terminated_length": 403.875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.233, "grad_norm": 0.07681547105312347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13424254.0, "reward": 0.7236729860305786, "reward_std": 0.3851258158683777, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25984376668930054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 530.25, "completions/mean_terminated_length": 530.25, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.2332, "grad_norm": 0.09846559166908264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13436450.0, "reward": 1.0671701431274414, "reward_std": 0.5064905881881714, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5718375444412231, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 1166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 588.0, "completions/mean_length": 512.375, "completions/mean_terminated_length": 478.2666931152344, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.2334, "grad_norm": 0.045062825083732605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13448360.0, "reward": 0.5434259176254272, "reward_std": 0.18308261036872864, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14934998750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8831249475479126, "step": 1167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 970.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 646.3125, "completions/mean_terminated_length": 646.3125, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.2336, "grad_norm": 0.13305819034576416, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13462877.0, "reward": 0.6409260034561157, "reward_std": 0.4588647782802582, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3443499803543091, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7925000190734863, "step": 1168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 464.0625, "completions/mean_terminated_length": 464.0625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.2338, "grad_norm": 0.13841600716114044, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13474318.0, "reward": 0.7981761693954468, "reward_std": 0.5055720806121826, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3463499844074249, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 670.625, "completions/mean_terminated_length": 620.1428833007812, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.234, "grad_norm": 0.06860657036304474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13488888.0, "reward": 0.630000650882721, "reward_std": 0.37154439091682434, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4475000202655792, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49375003576278687, "step": 1170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 400.5, "completions/mean_terminated_length": 400.5, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.2342, "grad_norm": 0.14474965631961823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13499440.0, "reward": 1.2863290309906006, "reward_std": 0.4711180031299591, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8226562738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 616.625, "completions/mean_terminated_length": 522.6154174804688, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2344, "grad_norm": 0.12386245280504227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13513674.0, "reward": 0.7339071035385132, "reward_std": 0.5681985020637512, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.46781253814697266, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.518750011920929, "step": 1172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 766.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 493.0625, "completions/mean_terminated_length": 493.0625, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.2346, "grad_norm": 0.038557231426239014, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13525755.0, "reward": 0.5114072561264038, "reward_std": 0.13691608607769012, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08531250059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.953125, "step": 1173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 577.3125, "completions/mean_terminated_length": 577.3125, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.2348, "grad_norm": 0.08766448497772217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13539200.0, "reward": 1.0421103239059448, "reward_std": 0.4569062888622284, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5842187404632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 1174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 412.125, "completions/mean_terminated_length": 412.125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.235, "grad_norm": 0.05706065893173218, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13550018.0, "reward": 1.4373953342437744, "reward_std": 0.24997177720069885, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.937287449836731, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9937499761581421, "step": 1175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 255.875, "completions/mean_terminated_length": 255.875, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.2352, "grad_norm": 0.15859444439411163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13557904.0, "reward": 1.1995916366577148, "reward_std": 0.45867830514907837, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7116812467575073, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 941.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 479.25, "completions/mean_terminated_length": 479.25, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.2354, "grad_norm": 0.03786096349358559, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13570404.0, "reward": 0.5314071178436279, "reward_std": 0.15703289210796356, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12531250715255737, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 1177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 472.75, "completions/mean_terminated_length": 472.75, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.2356, "grad_norm": 0.1053120419383049, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13582224.0, "reward": 0.8974853754043579, "reward_std": 0.48074647784233093, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41996875405311584, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 381.25, "completions/mean_terminated_length": 381.25, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2358, "grad_norm": 0.11998286843299866, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13592164.0, "reward": 1.0749012231826782, "reward_std": 0.4972156584262848, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5873000025749207, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 1179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 556.3125, "completions/mean_terminated_length": 556.3125, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.236, "grad_norm": 0.1270504593849182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13604681.0, "reward": 1.1270978450775146, "reward_std": 0.5178473591804504, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6916937828063965, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8706249594688416, "step": 1180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 517.375, "completions/mean_terminated_length": 483.60003662109375, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.2362, "grad_norm": 0.02241177298128605, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13616895.0, "reward": 0.5505476593971252, "reward_std": 0.16609059274196625, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16359373927116394, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.846875011920929, "step": 1181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 297.375, "completions/mean_terminated_length": 297.375, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.2364, "grad_norm": 0.054155800491571426, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13627109.0, "reward": 1.4396040439605713, "reward_std": 0.23756136000156403, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9417062997817993, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 529.875, "completions/mean_terminated_length": 529.875, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.2366, "grad_norm": 0.08113512396812439, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13639971.0, "reward": 0.9607820510864258, "reward_std": 0.3823724091053009, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6090624332427979, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.668749988079071, "step": 1183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 927.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 560.125, "completions/mean_terminated_length": 560.125, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.2368, "grad_norm": 0.07075761258602142, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13653093.0, "reward": 0.6258353590965271, "reward_std": 0.36696863174438477, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18916869163513184, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9522125124931335, "step": 1184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 314.75, "completions/mean_terminated_length": 314.75, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.237, "grad_norm": 0.1506858617067337, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13662001.0, "reward": 0.7839571833610535, "reward_std": 0.4266416132450104, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3179124891757965, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 430.875, "completions/mean_terminated_length": 430.875, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.2372, "grad_norm": 0.11172506958246231, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13672687.0, "reward": 0.9741420149803162, "reward_std": 0.4926345646381378, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5107812285423279, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 1186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 416.0, "completions/mean_terminated_length": 416.0, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.2374, "grad_norm": 0.06857436150312424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13683567.0, "reward": 0.6527325510978699, "reward_std": 0.3314143121242523, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18046249449253082, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9649999737739563, "step": 1187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 553.9375, "completions/mean_terminated_length": 553.9375, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.2376, "grad_norm": 0.10556410253047943, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13696830.0, "reward": 0.8083478808403015, "reward_std": 0.49931618571281433, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36669376492500305, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 859.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 626.1875, "completions/mean_terminated_length": 626.1875, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.2378, "grad_norm": 0.054709792137145996, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13710593.0, "reward": 0.726563572883606, "reward_std": 0.38389700651168823, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.265625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 370.5, "completions/mean_terminated_length": 370.5, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.238, "grad_norm": 0.06096750125288963, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13720649.0, "reward": 0.953750729560852, "reward_std": 0.23955151438713074, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7200000286102295, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.38749998807907104, "step": 1190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 402.375, "completions/mean_terminated_length": 402.375, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.2382, "grad_norm": 0.09867515414953232, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13731119.0, "reward": 1.2570793628692627, "reward_std": 0.43404924869537354, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7641562223434448, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 1191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 347.1875, "completions/mean_terminated_length": 347.1875, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.2384, "grad_norm": 0.11001522839069366, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13741122.0, "reward": 0.8107821345329285, "reward_std": 0.40625569224357605, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37156251072883606, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 740.25, "completions/mean_terminated_length": 674.7692260742188, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.2386, "grad_norm": 0.042674239724874496, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13757062.0, "reward": 0.5860943794250488, "reward_std": 0.3225540816783905, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3596875071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5562500357627869, "step": 1193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 494.4375, "completions/mean_terminated_length": 494.4375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.2388, "grad_norm": 0.04558250680565834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13768797.0, "reward": 0.752266526222229, "reward_std": 0.21107199788093567, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5045312643051147, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.578125, "step": 1194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 541.6875, "completions/mean_terminated_length": 541.6875, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.239, "grad_norm": 0.05843474715948105, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13781416.0, "reward": 0.6530478000640869, "reward_std": 0.2502993047237396, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2435937374830246, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 1195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 979.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 492.6875, "completions/mean_terminated_length": 492.6875, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.2392, "grad_norm": 0.07053028792142868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13793555.0, "reward": 0.7906257510185242, "reward_std": 0.3327648341655731, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.45625001192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.737500011920929, "step": 1196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 643.875, "completions/mean_terminated_length": 556.1538696289062, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.2394, "grad_norm": 0.11794499307870865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13808721.0, "reward": 0.4475788474082947, "reward_std": 0.3488141894340515, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20765623450279236, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.578125, "step": 1197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 261.3125, "completions/mean_terminated_length": 261.3125, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.2396, "grad_norm": 0.15972377359867096, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13816678.0, "reward": 1.197300910949707, "reward_std": 0.46255773305892944, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7071000337600708, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 625.125, "completions/mean_terminated_length": 625.125, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.2398, "grad_norm": 0.08556915819644928, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13830616.0, "reward": 0.780144989490509, "reward_std": 0.42949211597442627, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3102875053882599, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 1199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 518.625, "completions/mean_terminated_length": 518.625, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.24, "grad_norm": 0.045165207237005234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13843090.0, "reward": 1.3802978992462158, "reward_std": 0.3246031403541565, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855937719345093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 312.5625, "completions/mean_terminated_length": 312.5625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.2402, "grad_norm": 0.12398406118154526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13852395.0, "reward": 1.2897639274597168, "reward_std": 0.4666702151298523, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8295249938964844, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 1201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 397.6875, "completions/mean_terminated_length": 355.933349609375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.2404, "grad_norm": 0.05754292383790016, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13863254.0, "reward": 0.5659260153770447, "reward_std": 0.2833724021911621, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.131850004196167, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 1202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 290.3125, "completions/mean_terminated_length": 290.3125, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.2406, "grad_norm": 0.08215721696615219, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13871867.0, "reward": 1.3800759315490723, "reward_std": 0.3245161771774292, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851500153541565, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 475.75, "completions/mean_terminated_length": 475.75, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.2408, "grad_norm": 0.10469303280115128, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13883287.0, "reward": 1.079941749572754, "reward_std": 0.491471529006958, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5973812341690063, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 513.125, "completions/mean_terminated_length": 513.125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.241, "grad_norm": 0.23058637976646423, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13895785.0, "reward": 1.0053915977478027, "reward_std": 0.6269264817237854, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6357812881469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.871874988079071, "step": 1205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 251.6875, "completions/mean_terminated_length": 251.6875, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.2412, "grad_norm": 0.1565840244293213, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13906084.0, "reward": 1.0205978155136108, "reward_std": 0.49362149834632874, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5411937236785889, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518125057220459, "step": 1206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 469.9375, "completions/mean_terminated_length": 469.9375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.2414, "grad_norm": 0.11593254655599594, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13917315.0, "reward": 1.0758447647094727, "reward_std": 0.4954441487789154, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.589187502861023, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 441.625, "completions/mean_terminated_length": 441.625, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.2416, "grad_norm": 0.11952067166566849, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13928253.0, "reward": 0.9033479690551758, "reward_std": 0.47670888900756836, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4316937327384949, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1017.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 706.875, "completions/mean_terminated_length": 706.875, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.2418, "grad_norm": 0.09011256694793701, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13943755.0, "reward": 0.7639068365097046, "reward_std": 0.3841988444328308, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5903124809265137, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.43125003576278687, "step": 1209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 520.25, "completions/mean_terminated_length": 520.25, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.242, "grad_norm": 0.047251783311367035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13955967.0, "reward": 0.7914853096008301, "reward_std": 0.2291899174451828, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5829687714576721, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4781250059604645, "step": 1210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 381.1875, "completions/mean_terminated_length": 381.1875, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.2422, "grad_norm": 0.10472886264324188, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13965458.0, "reward": 0.8402323126792908, "reward_std": 0.45893919467926025, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.367962509393692, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 865.1875, "completions/mean_terminated_length": 769.9000244140625, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.2424, "grad_norm": 0.09712549299001694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13983621.0, "reward": 0.5575002431869507, "reward_std": 0.4503999948501587, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49000000953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.26250001788139343, "step": 1212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 390.875, "completions/mean_terminated_length": 390.875, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.2426, "grad_norm": 0.117934450507164, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13993955.0, "reward": 1.19303560256958, "reward_std": 0.503462553024292, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.6985687613487244, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8456250429153442, "step": 1213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 493.9375, "completions/mean_terminated_length": 493.9375, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.2428, "grad_norm": 0.049874741584062576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14006562.0, "reward": 0.6045072078704834, "reward_std": 0.23866796493530273, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1465124934911728, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 884.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 470.5, "completions/mean_terminated_length": 470.5, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.243, "grad_norm": 0.05695443972945213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14017786.0, "reward": 0.6181261539459229, "reward_std": 0.24107559025287628, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17374998331069946, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 1215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 702.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 530.125, "completions/mean_terminated_length": 530.125, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.2432, "grad_norm": 0.060814373195171356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14030124.0, "reward": 1.3762760162353516, "reward_std": 0.33626142144203186, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8775500059127808, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 1216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 280.0, "completions/max_terminated_length": 280.0, "completions/mean_length": 188.875, "completions/mean_terminated_length": 188.875, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.2434, "grad_norm": 0.0003006328479386866, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14036922.0, "reward": 1.4997137784957886, "reward_std": 0.0004830630496144295, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9994249939918518, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.98375004529953, "step": 1217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 409.5, "completions/mean_terminated_length": 409.5, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.2436, "grad_norm": 0.11446651071310043, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14047202.0, "reward": 0.7832044959068298, "reward_std": 0.4669035077095032, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37890625, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 1218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 339.375, "completions/mean_terminated_length": 339.375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.2438, "grad_norm": 0.1723451316356659, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14055992.0, "reward": 1.0142509937286377, "reward_std": 0.5010291934013367, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5285000205039978, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.966249942779541, "step": 1219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 365.75, "completions/mean_terminated_length": 365.75, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.244, "grad_norm": 0.07804129272699356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14065508.0, "reward": 0.5037513971328735, "reward_std": 0.14474397897720337, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07000000029802322, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 640.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 454.5625, "completions/mean_terminated_length": 454.5625, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.2442, "grad_norm": 0.18226158618927002, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14077149.0, "reward": 0.9655353426933289, "reward_std": 0.4860389232635498, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4935687780380249, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 306.125, "completions/mean_terminated_length": 306.125, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.2444, "grad_norm": 0.05981460586190224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14085967.0, "reward": 0.6425010561943054, "reward_std": 0.1881418526172638, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2849999964237213, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 1222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 482.375, "completions/mean_terminated_length": 482.375, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.2446, "grad_norm": 0.12987075746059418, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14097493.0, "reward": 1.1678822040557861, "reward_std": 0.5231107473373413, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7107625007629395, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9606125354766846, "step": 1223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 471.25, "completions/mean_terminated_length": 471.25, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.2448, "grad_norm": 0.11204739660024643, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14109353.0, "reward": 0.8342043161392212, "reward_std": 0.46344873309135437, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35590624809265137, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 1224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 348.1875, "completions/mean_terminated_length": 348.1875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.245, "grad_norm": 0.17977868020534515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14119564.0, "reward": 1.202494740486145, "reward_std": 0.4544380307197571, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7174875140190125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 414.875, "completions/mean_terminated_length": 414.875, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.2452, "grad_norm": 0.0757952332496643, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14129818.0, "reward": 0.6607009172439575, "reward_std": 0.27050551772117615, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.258899986743927, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 1226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 323.8125, "completions/mean_terminated_length": 323.8125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.2454, "grad_norm": 0.08579869568347931, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14138967.0, "reward": 0.6311730742454529, "reward_std": 0.2567606568336487, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19984374940395355, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843750357627869, "step": 1227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 278.5, "completions/mean_terminated_length": 278.5, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.2456, "grad_norm": 0.1906733214855194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14148015.0, "reward": 1.1354948282241821, "reward_std": 0.48509737849235535, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6459875106811523, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 1228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 469.3125, "completions/mean_terminated_length": 432.3333435058594, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.2458, "grad_norm": 0.14034941792488098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14159540.0, "reward": 0.892626166343689, "reward_std": 0.5798565745353699, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4727500081062317, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8525000214576721, "step": 1229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 266.3125, "completions/mean_terminated_length": 266.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.246, "grad_norm": 0.13504724204540253, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14167929.0, "reward": 1.320932388305664, "reward_std": 0.38310959935188293, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8293625116348267, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 1230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 810.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 549.0625, "completions/mean_terminated_length": 549.0625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2462, "grad_norm": 0.07949803024530411, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14180346.0, "reward": 1.2898446321487427, "reward_std": 0.31620001792907715, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8921874761581421, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 1231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 319.5, "completions/mean_terminated_length": 319.5, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.2464, "grad_norm": 0.11202637851238251, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14189730.0, "reward": 1.44025719165802, "reward_std": 0.23773549497127533, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9430124759674072, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 1232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 827.25, "completions/mean_terminated_length": 781.84619140625, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.2466, "grad_norm": 0.060076434165239334, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14206870.0, "reward": 0.6401569247245789, "reward_std": 0.3735620677471161, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4678124785423279, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.3999999761581421, "step": 1233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 786.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 524.0, "completions/mean_terminated_length": 524.0, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.2468, "grad_norm": 0.04445851221680641, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14219238.0, "reward": 1.4402296543121338, "reward_std": 0.2377282679080963, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9429562091827393, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831250309944153, "step": 1234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 554.8125, "completions/mean_terminated_length": 523.5333862304688, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.247, "grad_norm": 0.10717408359050751, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14232499.0, "reward": 0.6570071578025818, "reward_std": 0.3694584369659424, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2515124976634979, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8600000143051147, "step": 1235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 476.375, "completions/mean_terminated_length": 476.375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.2472, "grad_norm": 0.14868156611919403, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14244153.0, "reward": 1.0856664180755615, "reward_std": 0.5192329287528992, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6713312864303589, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8306249976158142, "step": 1236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 459.3125, "completions/mean_terminated_length": 459.3125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.2474, "grad_norm": 0.09963606297969818, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14255774.0, "reward": 0.6703915596008301, "reward_std": 0.3235630691051483, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21578124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 1237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 374.3125, "completions/mean_terminated_length": 374.3125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.2476, "grad_norm": 0.15837030112743378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14265411.0, "reward": 1.3800666332244873, "reward_std": 0.3245124816894531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 498.1875, "completions/mean_terminated_length": 498.1875, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.2478, "grad_norm": 0.130111426115036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14277590.0, "reward": 1.2019541263580322, "reward_std": 0.45483240485191345, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7164062261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 515.625, "completions/mean_terminated_length": 481.7333679199219, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.248, "grad_norm": 0.08264686167240143, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14289680.0, "reward": 0.6294539570808411, "reward_std": 0.24077720940113068, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3214062452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.703125, "step": 1240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 382.3125, "completions/mean_terminated_length": 382.3125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.2482, "grad_norm": 0.09333031624555588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14299813.0, "reward": 1.3767948150634766, "reward_std": 0.3328705430030823, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.878587543964386, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 416.8125, "completions/mean_terminated_length": 416.8125, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.2484, "grad_norm": 0.059271108359098434, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14310658.0, "reward": 0.6807758808135986, "reward_std": 0.27001726627349854, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2990500032901764, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8112500309944153, "step": 1242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 324.875, "completions/mean_terminated_length": 324.875, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.2486, "grad_norm": 0.1519552767276764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14320016.0, "reward": 1.1404447555541992, "reward_std": 0.47861602902412415, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6558874845504761, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 370.6875, "completions/mean_terminated_length": 370.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.2488, "grad_norm": 0.12849588692188263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14330603.0, "reward": 0.9063446521759033, "reward_std": 0.47469207644462585, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4376875162124634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 309.625, "completions/mean_terminated_length": 309.625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.249, "grad_norm": 0.10459200292825699, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14339317.0, "reward": 1.3185008764266968, "reward_std": 0.3884061574935913, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8244999647140503, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 1245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 525.125, "completions/mean_terminated_length": 525.125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.2492, "grad_norm": 0.02664531022310257, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14351927.0, "reward": 0.5877353549003601, "reward_std": 0.11522077023983002, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17546874284744263, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 1246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 226.75, "completions/mean_terminated_length": 226.75, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.2494, "grad_norm": 0.12181200832128525, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14358979.0, "reward": 0.8962509632110596, "reward_std": 0.3165830969810486, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6675000190734863, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5, "step": 1247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 449.4375, "completions/mean_terminated_length": 449.4375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.2496, "grad_norm": 0.093602254986763, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14370090.0, "reward": 0.7209072113037109, "reward_std": 0.38632962107658386, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2543124854564667, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 347.0, "completions/mean_terminated_length": 347.0, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.2498, "grad_norm": 0.1447477489709854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14379482.0, "reward": 1.194732666015625, "reward_std": 0.4675840735435486, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7019624710083008, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9862500429153442, "step": 1249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 361.8125, "completions/mean_terminated_length": 361.8125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.25, "grad_norm": 0.08156180381774902, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14389527.0, "reward": 0.6845259666442871, "reward_std": 0.28511062264442444, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.306549996137619, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7862499952316284, "step": 1250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 766.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 417.5625, "completions/mean_terminated_length": 417.5625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2502, "grad_norm": 0.03567403182387352, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14400096.0, "reward": 1.4350416660308838, "reward_std": 0.2261483073234558, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9325812458992004, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9356250166893005, "step": 1251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 668.75, "completions/mean_terminated_length": 668.75, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.2504, "grad_norm": 0.058630913496017456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14415532.0, "reward": 0.8967976570129395, "reward_std": 0.30666130781173706, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6685937643051147, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 1252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 671.125, "completions/mean_terminated_length": 647.6000366210938, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.2506, "grad_norm": 0.07181721925735474, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14430062.0, "reward": 0.5665633082389832, "reward_std": 0.26907840371131897, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2581250071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.765625, "step": 1253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 332.9375, "completions/mean_terminated_length": 332.9375, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.2508, "grad_norm": 0.125680610537529, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14438973.0, "reward": 0.8282818794250488, "reward_std": 0.2910654544830322, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5315625071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6937500238418579, "step": 1254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 464.125, "completions/mean_terminated_length": 464.125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.251, "grad_norm": 0.14297445118427277, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14450431.0, "reward": 1.0659105777740479, "reward_std": 0.5388718843460083, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6318187713623047, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9156250357627869, "step": 1255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 938.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 700.0, "completions/mean_terminated_length": 700.0, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.2512, "grad_norm": 0.0409482941031456, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14466335.0, "reward": 0.6607825756072998, "reward_std": 0.23327557742595673, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3215624988079071, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7125000357627869, "step": 1256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 604.125, "completions/mean_terminated_length": 576.1333618164062, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.2514, "grad_norm": 0.0847303569316864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14479921.0, "reward": 1.2273790836334229, "reward_std": 0.5016957521438599, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7672562599182129, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9043750166893005, "step": 1257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 449.75, "completions/mean_terminated_length": 449.75, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.2516, "grad_norm": 0.07647014409303665, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14491549.0, "reward": 0.927422285079956, "reward_std": 0.1672806292772293, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8548437356948853, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.24062499403953552, "step": 1258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 309.375, "completions/mean_terminated_length": 309.375, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.2518, "grad_norm": 0.17230825126171112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14500451.0, "reward": 0.8875511884689331, "reward_std": 0.4525335729122162, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4625999927520752, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 1259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 487.125, "completions/mean_terminated_length": 487.125, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.252, "grad_norm": 0.08659946918487549, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14512693.0, "reward": 1.374613642692566, "reward_std": 0.3414137661457062, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8742250204086304, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 1260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 259.625, "completions/mean_terminated_length": 259.625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.2522, "grad_norm": 0.1548703908920288, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14520687.0, "reward": 1.3149008750915527, "reward_std": 0.3504232168197632, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8797999620437622, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 1261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 332.0625, "completions/mean_terminated_length": 332.0625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.2524, "grad_norm": 0.14208726584911346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14529808.0, "reward": 0.8659353256225586, "reward_std": 0.44121500849723816, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4193687438964844, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 1262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 416.625, "completions/mean_terminated_length": 376.13336181640625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.2526, "grad_norm": 0.02976711466908455, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14540922.0, "reward": 0.5096040368080139, "reward_std": 0.1365329623222351, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08170624822378159, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 508.875, "completions/mean_terminated_length": 474.5333557128906, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.2528, "grad_norm": 0.10657130926847458, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14553288.0, "reward": 1.1334915161132812, "reward_std": 0.581112802028656, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7044812440872192, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 1264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 378.125, "completions/mean_terminated_length": 378.125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.253, "grad_norm": 0.009547989815473557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14563274.0, "reward": 0.5480228066444397, "reward_std": 0.015930647030472755, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09604375064373016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 1265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 737.8125, "completions/mean_terminated_length": 696.9285888671875, "completions/min_length": 551.0, "completions/min_terminated_length": 551.0, "epoch": 0.2532, "grad_norm": 0.09466254711151123, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14579095.0, "reward": 1.023353934288025, "reward_std": 0.5849083662033081, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6092062592506409, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8268749713897705, "step": 1266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 788.0, "completions/mean_terminated_length": 733.5385131835938, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.2534, "grad_norm": 0.07296251505613327, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14596391.0, "reward": 0.5805474519729614, "reward_std": 0.32050788402557373, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.34859374165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5843750238418579, "step": 1267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 896.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 762.8125, "completions/mean_terminated_length": 762.8125, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.2536, "grad_norm": 0.06233884021639824, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14612660.0, "reward": 0.8857038617134094, "reward_std": 0.33268263936042786, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5839061737060547, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.715624988079071, "step": 1268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 653.25, "completions/mean_terminated_length": 600.2857666015625, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.2538, "grad_norm": 0.11453751474618912, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14627400.0, "reward": 0.7425007820129395, "reward_std": 0.5661652684211731, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35999998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.828125, "step": 1269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 447.0625, "completions/mean_terminated_length": 447.0625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.254, "grad_norm": 0.17911724746227264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14638521.0, "reward": 1.109594702720642, "reward_std": 0.4687938094139099, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6566874980926514, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8949999809265137, "step": 1270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 395.8125, "completions/mean_terminated_length": 395.8125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.2542, "grad_norm": 0.14132378995418549, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14649206.0, "reward": 1.3802886009216309, "reward_std": 0.3245994448661804, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.885574996471405, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 765.8125, "completions/mean_terminated_length": 706.2307739257812, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.2544, "grad_norm": 0.0434705950319767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14665475.0, "reward": 0.48756638169288635, "reward_std": 0.2698957920074463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16263124346733093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7018749713897705, "step": 1272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 871.625, "completions/mean_terminated_length": 802.3636474609375, "completions/min_length": 688.0, "completions/min_terminated_length": 688.0, "epoch": 0.2546, "grad_norm": 0.0440543033182621, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14683869.0, "reward": 0.37998825311660767, "reward_std": 0.2688785791397095, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07247500121593475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6487500071525574, "step": 1273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 738.5, "completions/mean_terminated_length": 719.4666748046875, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.2548, "grad_norm": 0.050944529473781586, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14700037.0, "reward": 0.591638445854187, "reward_std": 0.23275165259838104, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2457749843597412, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7587499618530273, "step": 1274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 429.0625, "completions/mean_terminated_length": 429.0625, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.255, "grad_norm": 0.119426429271698, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14712102.0, "reward": 1.0130386352539062, "reward_std": 0.5023804306983948, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.526075005531311, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 492.5, "completions/mean_terminated_length": 492.5, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.2552, "grad_norm": 0.1113719493150711, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14724702.0, "reward": 0.8751543164253235, "reward_std": 0.47029587626457214, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.37530627846717834, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9243749976158142, "step": 1276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 469.6875, "completions/mean_terminated_length": 390.5000305175781, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.2554, "grad_norm": 0.036995649337768555, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14736041.0, "reward": 0.45882943272590637, "reward_std": 0.180649995803833, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.042656250298023224, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8531250357627869, "step": 1277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 447.25, "completions/mean_terminated_length": 447.25, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2556, "grad_norm": 0.16509254276752472, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14747453.0, "reward": 0.9646854996681213, "reward_std": 0.48709288239479065, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4918687641620636, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 1278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 400.0625, "completions/mean_terminated_length": 400.0625, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.2558, "grad_norm": 0.1254769116640091, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14757902.0, "reward": 1.2221072912216187, "reward_std": 0.43637850880622864, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7567124962806702, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 1279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 345.0, "completions/max_terminated_length": 345.0, "completions/mean_length": 217.625, "completions/mean_terminated_length": 217.625, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.256, "grad_norm": 0.2293720543384552, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14765304.0, "reward": 0.835919976234436, "reward_std": 0.46249619126319885, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3593375086784363, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 1280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 323.6875, "completions/mean_terminated_length": 323.6875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.2562, "grad_norm": 0.12673938274383545, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14774259.0, "reward": 1.255620002746582, "reward_std": 0.4362114369869232, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7612375020980835, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 963.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 611.25, "completions/mean_terminated_length": 611.25, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.2564, "grad_norm": 0.11649274080991745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14789479.0, "reward": 0.8505321741104126, "reward_std": 0.4827060401439667, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45106250047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 1282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 523.8125, "completions/mean_terminated_length": 523.8125, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.2566, "grad_norm": 0.11246608197689056, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14801732.0, "reward": 1.313682198524475, "reward_std": 0.3987954556941986, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8148625493049622, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 347.25, "completions/mean_terminated_length": 347.25, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.2568, "grad_norm": 0.14164187014102936, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14811368.0, "reward": 0.8903888463973999, "reward_std": 0.487703412771225, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4057750105857849, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 1284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 668.0625, "completions/mean_terminated_length": 668.0625, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.257, "grad_norm": 0.036015551537275314, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14826809.0, "reward": 0.5972260236740112, "reward_std": 0.2410115897655487, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1319499909877777, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 636.4375, "completions/mean_terminated_length": 581.0714721679688, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.2572, "grad_norm": 0.031733438372612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14841328.0, "reward": 0.4900009036064148, "reward_std": 0.19865305721759796, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10499999672174454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 1286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 253.6875, "completions/mean_terminated_length": 253.6875, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.2574, "grad_norm": 0.10936357825994492, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14849723.0, "reward": 1.377922773361206, "reward_std": 0.33132466673851013, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8808437585830688, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 460.9375, "completions/mean_terminated_length": 460.9375, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2576, "grad_norm": 0.006570653989911079, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14861018.0, "reward": 0.5256761908531189, "reward_std": 0.02156466618180275, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05134999752044678, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 438.8125, "completions/mean_terminated_length": 399.8000183105469, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.2578, "grad_norm": 0.10816384106874466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14872023.0, "reward": 0.7031134366989136, "reward_std": 0.4213307797908783, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2812249958515167, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8799999952316284, "step": 1289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 320.5, "completions/mean_terminated_length": 320.5, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.258, "grad_norm": 0.0887899324297905, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14881007.0, "reward": 1.3839702606201172, "reward_std": 0.31712159514427185, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8929374814033508, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.987500011920929, "step": 1290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 299.0, "completions/max_terminated_length": 299.0, "completions/mean_length": 213.4375, "completions/mean_terminated_length": 213.4375, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.2582, "grad_norm": 0.2135694921016693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14888438.0, "reward": 0.8192765116691589, "reward_std": 0.47412294149398804, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3260499835014343, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9925000071525574, "step": 1291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 379.1875, "completions/mean_terminated_length": 379.1875, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.2584, "grad_norm": 0.16383782029151917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14898249.0, "reward": 1.2562353610992432, "reward_std": 0.43357715010643005, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7624687552452087, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 431.375, "completions/mean_terminated_length": 391.86669921875, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.2586, "grad_norm": 0.06197986751794815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14909631.0, "reward": 1.4053666591644287, "reward_std": 0.3747647702693939, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9357312321662903, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 1293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 852.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 633.3125, "completions/mean_terminated_length": 633.3125, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.2588, "grad_norm": 0.09790948778390884, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14924100.0, "reward": 1.097266435623169, "reward_std": 0.4790334701538086, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.757031261920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.746874988079071, "step": 1294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 323.25, "completions/mean_terminated_length": 323.25, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.259, "grad_norm": 0.1260974109172821, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14933960.0, "reward": 1.1979291439056396, "reward_std": 0.4615679383277893, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7083562612533569, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 737.625, "completions/mean_terminated_length": 696.7142944335938, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.2592, "grad_norm": 0.05552147701382637, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14950290.0, "reward": 0.750234842300415, "reward_std": 0.3716994822025299, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5629687309265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 1296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 624.5, "completions/mean_terminated_length": 624.5, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.2594, "grad_norm": 0.10272443294525146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14964378.0, "reward": 1.011563777923584, "reward_std": 0.503479540348053, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5231249928474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 1297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 452.1875, "completions/mean_terminated_length": 452.1875, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2596, "grad_norm": 0.13943733274936676, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14976205.0, "reward": 1.0808541774749756, "reward_std": 0.48979753255844116, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.599206268787384, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 377.4375, "completions/mean_terminated_length": 377.4375, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.2598, "grad_norm": 0.06512017548084259, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14986164.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 583.1875, "completions/mean_terminated_length": 553.800048828125, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.26, "grad_norm": 0.08856162428855896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14999335.0, "reward": 1.2270324230194092, "reward_std": 0.5014963746070862, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7665624618530273, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 502.25, "completions/mean_terminated_length": 502.25, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.2602, "grad_norm": 0.08762194961309433, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15011627.0, "reward": 0.8025009632110596, "reward_std": 0.4177294075489044, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35500001907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 1301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 425.4375, "completions/mean_terminated_length": 425.4375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.2604, "grad_norm": 0.009439732879400253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15022498.0, "reward": 0.5593758821487427, "reward_std": 0.04249993711709976, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11874999105930328, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 1302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 371.625, "completions/mean_terminated_length": 371.625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.2606, "grad_norm": 0.13810613751411438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15032364.0, "reward": 1.0807665586471558, "reward_std": 0.4897176921367645, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5990312695503235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 279.0, "completions/max_terminated_length": 279.0, "completions/mean_length": 223.3125, "completions/mean_terminated_length": 223.3125, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.2608, "grad_norm": 0.28844913840293884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15039521.0, "reward": 1.1905823945999146, "reward_std": 0.47393083572387695, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6936625242233276, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9912500381469727, "step": 1304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 327.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 260.3125, "completions/mean_terminated_length": 260.3125, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.261, "grad_norm": 0.205503910779953, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15047398.0, "reward": 0.8287419080734253, "reward_std": 0.46728700399398804, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34498125314712524, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 1305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 545.25, "completions/mean_terminated_length": 513.3333740234375, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.2612, "grad_norm": 0.11369454115629196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15059946.0, "reward": 1.2807135581970215, "reward_std": 0.481706440448761, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.811424970626831, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 1306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 267.1875, "completions/mean_terminated_length": 267.1875, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2614, "grad_norm": 0.27601057291030884, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15068477.0, "reward": 1.345997929573059, "reward_std": 0.4301486611366272, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.879493772983551, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 485.375, "completions/mean_terminated_length": 485.375, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.2616, "grad_norm": 0.14331525564193726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15080051.0, "reward": 0.8436571955680847, "reward_std": 0.45585039258003235, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.374812513589859, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 1308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 369.9375, "completions/mean_terminated_length": 369.9375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.2618, "grad_norm": 0.12908025085926056, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15090082.0, "reward": 0.9019541144371033, "reward_std": 0.4775867462158203, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42890626192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 945.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 768.25, "completions/mean_terminated_length": 768.25, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.262, "grad_norm": 0.05806176736950874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15106454.0, "reward": 0.9500788450241089, "reward_std": 0.36439529061317444, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5876562595367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.640625, "step": 1310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 708.9375, "completions/mean_terminated_length": 687.933349609375, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2622, "grad_norm": 0.038441672921180725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15121717.0, "reward": 0.5667980909347534, "reward_std": 0.2064247876405716, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19609373807907104, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.815625011920929, "step": 1311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 337.375, "completions/mean_terminated_length": 337.375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2624, "grad_norm": 0.04162507504224777, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15131291.0, "reward": 0.6349226832389832, "reward_std": 0.125759556889534, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.269843727350235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8343750238418579, "step": 1312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 851.5625, "completions/mean_terminated_length": 748.1000366210938, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.2626, "grad_norm": 0.07001104950904846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15149492.0, "reward": 0.4376567006111145, "reward_std": 0.3676479160785675, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.250312477350235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.46875, "step": 1313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 862.6875, "completions/mean_terminated_length": 789.3636474609375, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.2628, "grad_norm": 0.10267315804958344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15167471.0, "reward": 0.7141286134719849, "reward_std": 0.5869836211204529, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.49075624346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5143749713897705, "step": 1314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 336.0, "completions/max_terminated_length": 336.0, "completions/mean_length": 296.0625, "completions/mean_terminated_length": 296.0625, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.263, "grad_norm": 0.007072148844599724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15175792.0, "reward": 0.5390509366989136, "reward_std": 0.01735185645520687, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07809999585151672, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 410.75, "completions/mean_terminated_length": 410.75, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.2632, "grad_norm": 0.092979796230793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15186140.0, "reward": 1.26442289352417, "reward_std": 0.4199773371219635, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7788437604904175, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 594.875, "completions/mean_terminated_length": 533.5714721679688, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.2634, "grad_norm": 0.13060404360294342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15199450.0, "reward": 0.9525793790817261, "reward_std": 0.590461015701294, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5301562547683716, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 1317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 976.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 489.875, "completions/mean_terminated_length": 489.875, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.2636, "grad_norm": 0.15297618508338928, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15211496.0, "reward": 1.1783602237701416, "reward_std": 0.37846365571022034, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7317187190055847, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.684374988079071, "step": 1318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 455.3125, "completions/mean_terminated_length": 455.3125, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.2638, "grad_norm": 0.11141139268875122, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15223021.0, "reward": 0.7170135974884033, "reward_std": 0.38860806822776794, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24652500450611115, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 516.375, "completions/mean_terminated_length": 516.375, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.264, "grad_norm": 0.10525023937225342, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15235139.0, "reward": 1.1345982551574707, "reward_std": 0.4870675206184387, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6441937685012817, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750619888306, "step": 1320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 422.1875, "completions/mean_terminated_length": 422.1875, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.2642, "grad_norm": 0.14227721095085144, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15247078.0, "reward": 1.255666732788086, "reward_std": 0.43623921275138855, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7613312602043152, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 1321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 568.875, "completions/mean_terminated_length": 538.5333862304688, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.2644, "grad_norm": 0.1079266294836998, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15260852.0, "reward": 1.105494737625122, "reward_std": 0.5403302907943726, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6484875082969666, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 1322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 407.3125, "completions/mean_terminated_length": 366.20001220703125, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.2646, "grad_norm": 0.09650681912899017, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15271849.0, "reward": 0.6384382247924805, "reward_std": 0.24773284792900085, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.33937492966651917, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6688125133514404, "step": 1323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 291.8125, "completions/mean_terminated_length": 291.8125, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.2648, "grad_norm": 0.0002734007139224559, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15280518.0, "reward": 1.4990978240966797, "reward_std": 0.000533094338607043, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981937408447266, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 440.375, "completions/mean_terminated_length": 440.375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.265, "grad_norm": 0.09668698161840439, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15291820.0, "reward": 0.7295323014259338, "reward_std": 0.3827483355998993, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27156248688697815, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 789.9375, "completions/mean_terminated_length": 735.923095703125, "completions/min_length": 594.0, "completions/min_terminated_length": 594.0, "epoch": 0.2652, "grad_norm": 0.11792431771755219, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15308939.0, "reward": 0.7024601697921753, "reward_std": 0.596889853477478, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34241876006126404, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7800000309944153, "step": 1326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 330.6875, "completions/mean_terminated_length": 330.6875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.2654, "grad_norm": 0.14211030304431915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15318134.0, "reward": 1.082891583442688, "reward_std": 0.4865175485610962, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6032812595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 1327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 827.8125, "completions/mean_terminated_length": 782.5385131835938, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.2656, "grad_norm": 0.028057195246219635, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15335763.0, "reward": 0.4707821011543274, "reward_std": 0.25115838646888733, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1290625035762787, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.737500011920929, "step": 1328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 315.375, "completions/mean_terminated_length": 315.375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2658, "grad_norm": 0.0001949396828422323, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15345449.0, "reward": 1.4992167949676514, "reward_std": 0.0006213470478542149, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9984312653541565, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 1329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 703.25, "completions/mean_terminated_length": 681.86669921875, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.266, "grad_norm": 0.07889920473098755, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15361149.0, "reward": 1.0862383842468262, "reward_std": 0.48420047760009766, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7349749803543091, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7362500429153442, "step": 1330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 558.3125, "completions/mean_terminated_length": 558.3125, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.2662, "grad_norm": 0.00166740280110389, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15374946.0, "reward": 0.5469415187835693, "reward_std": 0.007237497717142105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09388124942779541, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 524.75, "completions/mean_terminated_length": 524.75, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.2664, "grad_norm": 0.053187254816293716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15388494.0, "reward": 0.5935137271881104, "reward_std": 0.2424117475748062, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12452499568462372, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9649999737739563, "step": 1332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 595.5, "completions/mean_terminated_length": 534.2857666015625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.2666, "grad_norm": 0.11101709306240082, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15401782.0, "reward": 0.5312509536743164, "reward_std": 0.25674745440483093, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1875, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.731249988079071, "step": 1333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 651.3125, "completions/mean_terminated_length": 651.3125, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.2668, "grad_norm": 0.08676639944314957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15416731.0, "reward": 0.9538856744766235, "reward_std": 0.4975210130214691, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47026875615119934, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781249761581421, "step": 1334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 472.375, "completions/mean_terminated_length": 472.375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.267, "grad_norm": 0.08682925254106522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15427857.0, "reward": 1.3211292028427124, "reward_std": 0.3832072913646698, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8297562599182129, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 938.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 656.3125, "completions/mean_terminated_length": 656.3125, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.2672, "grad_norm": 0.05229956656694412, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15442710.0, "reward": 0.8266287446022034, "reward_std": 0.22726552188396454, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6532562375068665, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.40187498927116394, "step": 1336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 763.625, "completions/mean_terminated_length": 746.2667236328125, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.2674, "grad_norm": 0.06691381335258484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15459984.0, "reward": 0.6425759196281433, "reward_std": 0.38006269931793213, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3476499915122986, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6899999380111694, "step": 1337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 470.0625, "completions/mean_terminated_length": 470.0625, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.2676, "grad_norm": 0.004868897143751383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15471601.0, "reward": 0.5347605347633362, "reward_std": 0.02161562442779541, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06951874494552612, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 629.6875, "completions/mean_terminated_length": 573.357177734375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2678, "grad_norm": 0.06587209552526474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15485724.0, "reward": 0.7051568031311035, "reward_std": 0.3284663259983063, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5353125333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 1339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 374.6875, "completions/mean_terminated_length": 374.6875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.268, "grad_norm": 0.20972150564193726, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15496135.0, "reward": 0.9185106158256531, "reward_std": 0.544036328792572, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4620187282562256, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 717.3125, "completions/mean_terminated_length": 673.5, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.2682, "grad_norm": 0.07367003709077835, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15511372.0, "reward": 0.5729570388793945, "reward_std": 0.27490177750587463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2709124982357025, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6924999952316284, "step": 1341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 408.6875, "completions/mean_terminated_length": 408.6875, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.2684, "grad_norm": 0.0001284221070818603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15522199.0, "reward": 1.4989603757858276, "reward_std": 0.0004522755916696042, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979187250137329, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 461.8125, "completions/mean_terminated_length": 461.8125, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.2686, "grad_norm": 0.004798999521881342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15533764.0, "reward": 0.5408978462219238, "reward_std": 0.014219691976904869, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08179374784231186, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 386.9375, "completions/mean_terminated_length": 386.9375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.2688, "grad_norm": 0.17460928857326508, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15543795.0, "reward": 1.1347417831420898, "reward_std": 0.48585331439971924, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6444812417030334, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 405.625, "completions/mean_terminated_length": 405.625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.269, "grad_norm": 0.1088893935084343, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15554269.0, "reward": 1.3768635988235474, "reward_std": 0.3328975439071655, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.878724992275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 646.625, "completions/mean_terminated_length": 646.625, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.2692, "grad_norm": 0.055685270577669144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15568327.0, "reward": 0.6535793542861938, "reward_std": 0.3308374881744385, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1821562498807907, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 483.3125, "completions/mean_terminated_length": 483.3125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.2694, "grad_norm": 0.10817261785268784, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15579884.0, "reward": 0.9338290691375732, "reward_std": 0.44595393538475037, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49265626072883606, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.840624988079071, "step": 1347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 285.125, "completions/mean_terminated_length": 285.125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2696, "grad_norm": 0.20157890021800995, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15588414.0, "reward": 1.1949759721755981, "reward_std": 0.4649538993835449, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7024500370025635, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 661.3125, "completions/mean_terminated_length": 637.1333618164062, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.2698, "grad_norm": 0.05711514130234718, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15603043.0, "reward": 0.7196884155273438, "reward_std": 0.37296441197395325, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.37687498331069946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 1349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 603.5625, "completions/mean_terminated_length": 603.5625, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.27, "grad_norm": 0.07574190944433212, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15616636.0, "reward": 0.8044542074203491, "reward_std": 0.4000324010848999, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35890623927116394, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "step": 1350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 862.0, "completions/mean_length": 551.4375, "completions/mean_terminated_length": 519.933349609375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.2702, "grad_norm": 0.1143791452050209, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15629507.0, "reward": 0.6864855289459229, "reward_std": 0.4257636070251465, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24796873331069946, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.903124988079071, "step": 1351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 596.0625, "completions/mean_terminated_length": 596.0625, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.2704, "grad_norm": 0.12190531194210052, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15643396.0, "reward": 1.0737512111663818, "reward_std": 0.4267558157444, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7100000381469727, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 1352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 400.6875, "completions/mean_terminated_length": 400.6875, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.2706, "grad_norm": 0.09244948625564575, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15653695.0, "reward": 1.3141759634017944, "reward_std": 0.39900752902030945, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.815850019454956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 1353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 720.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 429.0625, "completions/mean_terminated_length": 429.0625, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.2708, "grad_norm": 0.027714423835277557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15665728.0, "reward": 0.6471883654594421, "reward_std": 0.10096552968025208, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2943750023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 1354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 300.75, "completions/mean_terminated_length": 300.75, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.271, "grad_norm": 0.12320097535848618, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15674364.0, "reward": 1.315535545349121, "reward_std": 0.39502283930778503, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.818568766117096, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 318.375, "completions/mean_terminated_length": 318.375, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.2712, "grad_norm": 0.18315456807613373, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15683570.0, "reward": 0.6608353853225708, "reward_std": 0.3274187743663788, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19666874408721924, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 547.25, "completions/mean_terminated_length": 515.4666748046875, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.2714, "grad_norm": 0.10736732184886932, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15696294.0, "reward": 1.1080477237701416, "reward_std": 0.5364318490028381, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6535937786102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 1357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 683.75, "completions/mean_terminated_length": 661.0667114257812, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.2716, "grad_norm": 0.07511603832244873, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15711618.0, "reward": 0.4849107265472412, "reward_std": 0.22415122389793396, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09481875598430634, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8018749952316284, "step": 1358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 553.0, "completions/mean_terminated_length": 553.0, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.2718, "grad_norm": 0.037484098225831985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15724322.0, "reward": 0.7555474638938904, "reward_std": 0.18325097858905792, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5110937356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.628125011920929, "step": 1359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 373.375, "completions/mean_terminated_length": 373.375, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.272, "grad_norm": 0.09076009690761566, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15734056.0, "reward": 0.6281633973121643, "reward_std": 0.2588122487068176, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19382500648498535, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 1360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 324.0625, "completions/mean_terminated_length": 324.0625, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.2722, "grad_norm": 0.1894589513540268, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15742841.0, "reward": 0.8454697132110596, "reward_std": 0.4552173614501953, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37843751907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 500.9375, "completions/mean_terminated_length": 466.0666809082031, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.2724, "grad_norm": 0.11482922732830048, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15755464.0, "reward": 0.8948444128036499, "reward_std": 0.4793444275856018, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5396875143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 1362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 735.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 548.125, "completions/mean_terminated_length": 548.125, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.2726, "grad_norm": 0.12182220816612244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15768122.0, "reward": 0.8707040548324585, "reward_std": 0.5201067328453064, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42890626192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 1363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 544.625, "completions/mean_terminated_length": 544.625, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.2728, "grad_norm": 0.1037215143442154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15780580.0, "reward": 0.9123040437698364, "reward_std": 0.43506649136543274, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5121062397956848, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8381249904632568, "step": 1364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 285.3125, "completions/mean_terminated_length": 285.3125, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.273, "grad_norm": 0.1899789571762085, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15789065.0, "reward": 1.0689074993133545, "reward_std": 0.5043798089027405, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5753124952316284, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781093597412109, "step": 1365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 679.5, "completions/mean_terminated_length": 564.6666870117188, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.2732, "grad_norm": 0.11101873219013214, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15804129.0, "reward": 0.820859968662262, "reward_std": 0.6177787184715271, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5167187452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6281249523162842, "step": 1366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 258.4375, "completions/mean_terminated_length": 258.4375, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.2734, "grad_norm": 0.13151633739471436, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15811576.0, "reward": 0.721479058265686, "reward_std": 0.3325403332710266, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.31795623898506165, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 1367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 485.625, "completions/mean_terminated_length": 485.625, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.2736, "grad_norm": 0.07002179324626923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15823474.0, "reward": 1.3771761655807495, "reward_std": 0.3330181837081909, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8793500065803528, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 1368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 587.0, "completions/mean_terminated_length": 557.86669921875, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.2738, "grad_norm": 0.09941082447767258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15836834.0, "reward": 1.0067977905273438, "reward_std": 0.5337310433387756, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5760937333106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 1369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 421.5, "completions/mean_terminated_length": 421.5, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.274, "grad_norm": 0.10400784760713577, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15847194.0, "reward": 0.8438822031021118, "reward_std": 0.4561940133571625, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3752624988555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 175.0, "completions/max_terminated_length": 175.0, "completions/mean_length": 148.5625, "completions/mean_terminated_length": 148.5625, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 0.2742, "grad_norm": 0.17964574694633484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15853427.0, "reward": 0.6354419589042664, "reward_std": 0.3378104269504547, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14588123559951782, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9881249666213989, "step": 1371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 355.0, "completions/mean_terminated_length": 355.0, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2744, "grad_norm": 0.135416641831398, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15863747.0, "reward": 1.3157696723937988, "reward_std": 0.39349958300590515, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8190374970436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 437.1875, "completions/mean_terminated_length": 398.0666809082031, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.2746, "grad_norm": 0.11478989571332932, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15875398.0, "reward": 0.7794542908668518, "reward_std": 0.42625531554222107, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.433906227350235, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6781250238418579, "step": 1373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 584.75, "completions/mean_terminated_length": 584.75, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.2748, "grad_norm": 0.08728893101215363, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15889602.0, "reward": 0.7811727523803711, "reward_std": 0.37387815117836, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3748437464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 1374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 441.5, "completions/mean_terminated_length": 441.5, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.275, "grad_norm": 0.06411046534776688, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15900666.0, "reward": 0.6974226236343384, "reward_std": 0.1830538511276245, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3948437571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.715624988079071, "step": 1375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 395.5625, "completions/mean_terminated_length": 395.5625, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.2752, "grad_norm": 0.13456867635250092, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15911059.0, "reward": 0.8926759958267212, "reward_std": 0.48605072498321533, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4103499948978424, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 1376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 367.0, "completions/mean_terminated_length": 367.0, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.2754, "grad_norm": 0.15729016065597534, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15921683.0, "reward": 1.207576036453247, "reward_std": 0.4460582733154297, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.727649986743927, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 1377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 800.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 586.25, "completions/mean_terminated_length": 586.25, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.2756, "grad_norm": 0.0813668817281723, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15935927.0, "reward": 0.6732822060585022, "reward_std": 0.3226090967655182, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2215624898672104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 1378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 713.3125, "completions/mean_terminated_length": 692.6000366210938, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.2758, "grad_norm": 0.044356729835271835, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15951196.0, "reward": 0.9364066123962402, "reward_std": 0.3270752429962158, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8103125095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.38750001788139343, "step": 1379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 578.0, "completions/mean_terminated_length": 548.2667236328125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.276, "grad_norm": 0.08835949003696442, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15964156.0, "reward": 0.885541558265686, "reward_std": 0.44023439288139343, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5835812091827393, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6643750071525574, "step": 1380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 542.625, "completions/mean_terminated_length": 542.625, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.2762, "grad_norm": 0.03644144907593727, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15977990.0, "reward": 0.741485059261322, "reward_std": 0.1206708550453186, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4829687476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6906249523162842, "step": 1381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 865.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 521.4375, "completions/mean_terminated_length": 521.4375, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.2764, "grad_norm": 0.0818939134478569, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15990445.0, "reward": 0.8582038879394531, "reward_std": 0.37446752190589905, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5289062261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7156249284744263, "step": 1382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 290.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 205.3125, "completions/mean_terminated_length": 205.3125, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.2766, "grad_norm": 0.18528501689434052, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15997474.0, "reward": 1.194732427597046, "reward_std": 0.4675559997558594, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7019624710083008, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 1383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 333.875, "completions/mean_terminated_length": 333.875, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.2768, "grad_norm": 0.06897449493408203, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16006752.0, "reward": 0.5935573577880859, "reward_std": 0.24221499264240265, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12461249530315399, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 558.6875, "completions/mean_terminated_length": 558.6875, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.277, "grad_norm": 0.11795654147863388, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16019851.0, "reward": 0.8066415786743164, "reward_std": 0.42050108313560486, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.36328125, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 1385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 474.0, "completions/mean_terminated_length": 437.3333435058594, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2772, "grad_norm": 0.08162669837474823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16031515.0, "reward": 0.7095320820808411, "reward_std": 0.38351336121559143, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3565624952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.75, "step": 1386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 885.25, "completions/mean_terminated_length": 777.3333129882812, "completions/min_length": 643.0, "completions/min_terminated_length": 643.0, "epoch": 0.2774, "grad_norm": 0.10538582503795624, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16049551.0, "reward": 0.4272661805152893, "reward_std": 0.5553101301193237, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2920312285423279, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.42500001192092896, "step": 1387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 448.5, "completions/mean_terminated_length": 448.5, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.2776, "grad_norm": 0.1086459755897522, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16060823.0, "reward": 1.1931697130203247, "reward_std": 0.43086159229278564, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7613375186920166, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8612500429153442, "step": 1388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 541.8125, "completions/mean_terminated_length": 541.8125, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.2778, "grad_norm": 0.00016311235958710313, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16073732.0, "reward": 1.499213695526123, "reward_std": 0.0006171901477500796, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9984250068664551, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 721.75, "completions/mean_terminated_length": 721.75, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.278, "grad_norm": 0.059714894741773605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16089072.0, "reward": 0.8036727905273438, "reward_std": 0.3156992495059967, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.48234376311302185, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.753125011920929, "step": 1390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 425.5, "completions/mean_terminated_length": 425.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.2782, "grad_norm": 0.14875979721546173, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16099720.0, "reward": 1.2614572048187256, "reward_std": 0.42497608065605164, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7729125022888184, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 653.75, "completions/mean_terminated_length": 629.0667114257812, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.2784, "grad_norm": 0.09281709790229797, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16115060.0, "reward": 0.7089823484420776, "reward_std": 0.43266561627388, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2929624915122986, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8525000214576721, "step": 1392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 735.9375, "completions/mean_terminated_length": 639.9166870117188, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.2786, "grad_norm": 0.05402253195643425, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16130787.0, "reward": 0.4503132700920105, "reward_std": 0.29742658138275146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15062499046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6437499523162842, "step": 1393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 359.4375, "completions/mean_terminated_length": 359.4375, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.2788, "grad_norm": 0.15853722393512726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16140058.0, "reward": 0.851394772529602, "reward_std": 0.45372509956359863, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39028748869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 1394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 413.3125, "completions/mean_terminated_length": 413.3125, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.279, "grad_norm": 0.14845874905586243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16150495.0, "reward": 1.2559170722961426, "reward_std": 0.4363882839679718, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7618312239646912, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 1395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 838.25, "completions/mean_terminated_length": 825.86669921875, "completions/min_length": 616.0, "completions/min_terminated_length": 616.0, "epoch": 0.2792, "grad_norm": 0.05364162474870682, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16168211.0, "reward": 0.6103914976119995, "reward_std": 0.300577849149704, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.34578123688697815, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.690625011920929, "step": 1396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 743.375, "completions/mean_terminated_length": 724.6666870117188, "completions/min_length": 528.0, "completions/min_terminated_length": 528.0, "epoch": 0.2794, "grad_norm": 0.10553734749555588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16185689.0, "reward": 1.0779695510864258, "reward_std": 0.4918334484100342, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7184374332427979, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.731249988079071, "step": 1397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 436.625, "completions/mean_terminated_length": 436.625, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.2796, "grad_norm": 0.13974250853061676, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16196979.0, "reward": 0.8992478251457214, "reward_std": 0.48037856817245483, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42349374294281006, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 1398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 377.0625, "completions/mean_terminated_length": 377.0625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.2798, "grad_norm": 0.0799049586057663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16206692.0, "reward": 1.3754353523254395, "reward_std": 0.33788955211639404, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8758687973022461, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 336.0, "completions/max_terminated_length": 336.0, "completions/mean_length": 163.125, "completions/mean_terminated_length": 163.125, "completions/min_length": 108.0, "completions/min_terminated_length": 108.0, "epoch": 0.28, "grad_norm": 0.17068466544151306, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16213030.0, "reward": 1.4382543563842773, "reward_std": 0.23726579546928406, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9390062093734741, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 371.0, "completions/max_terminated_length": 371.0, "completions/mean_length": 249.6875, "completions/mean_terminated_length": 249.6875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.2802, "grad_norm": 0.19228845834732056, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16220753.0, "reward": 1.3775701522827148, "reward_std": 0.333172082901001, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8801375031471252, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "step": 1401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 586.0625, "completions/mean_terminated_length": 556.86669921875, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.2804, "grad_norm": 0.0936536118388176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16233938.0, "reward": 1.0091387033462524, "reward_std": 0.5381900668144226, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5807750225067139, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 1402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 401.9375, "completions/mean_terminated_length": 401.9375, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.2806, "grad_norm": 0.05784953013062477, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16245473.0, "reward": 0.5849950909614563, "reward_std": 0.24477897584438324, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1074874997138977, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 828.0625, "completions/mean_terminated_length": 782.84619140625, "completions/min_length": 563.0, "completions/min_terminated_length": 563.0, "epoch": 0.2808, "grad_norm": 0.06085929647088051, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16262754.0, "reward": 0.596406877040863, "reward_std": 0.3829342722892761, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3178125023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.643750011920929, "step": 1404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 587.25, "completions/mean_terminated_length": 558.1333618164062, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.281, "grad_norm": 0.1173495203256607, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16276070.0, "reward": 1.051797866821289, "reward_std": 0.540323793888092, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6035937666893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 1405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 760.0625, "completions/mean_terminated_length": 601.7000122070312, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.2812, "grad_norm": 0.08136496692895889, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16291831.0, "reward": 0.45459455251693726, "reward_std": 0.4816219210624695, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15918749570846558, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6075000166893005, "step": 1406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 349.875, "completions/mean_terminated_length": 349.875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.2814, "grad_norm": 0.16398562490940094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16301733.0, "reward": 1.3313947916030884, "reward_std": 0.3619644045829773, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8502874970436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 1407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 940.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 554.25, "completions/mean_terminated_length": 554.25, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.2816, "grad_norm": 0.10329195111989975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16314457.0, "reward": 1.110457181930542, "reward_std": 0.45859605073928833, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6584124565124512, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 1408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 363.0, "completions/max_terminated_length": 363.0, "completions/mean_length": 296.875, "completions/mean_terminated_length": 296.875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2818, "grad_norm": 0.00016448748647235334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16322855.0, "reward": 1.4989041090011597, "reward_std": 0.0004186927108094096, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9978062510490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 414.3125, "completions/mean_terminated_length": 414.3125, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.282, "grad_norm": 0.04229872673749924, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16333324.0, "reward": 0.7008507251739502, "reward_std": 0.13743625581264496, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4016999900341034, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.75, "step": 1410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 520.1875, "completions/mean_terminated_length": 520.1875, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2822, "grad_norm": 0.09519968181848526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16345599.0, "reward": 1.341200828552246, "reward_std": 0.24248774349689484, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.9948999881744385, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7000000476837158, "step": 1411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 294.75, "completions/mean_terminated_length": 294.75, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2824, "grad_norm": 0.2256140559911728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16354155.0, "reward": 1.4051446914672852, "reward_std": 0.3747048079967499, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9352874755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 892.0625, "completions/mean_terminated_length": 848.0833740234375, "completions/min_length": 693.0, "completions/min_terminated_length": 693.0, "epoch": 0.2826, "grad_norm": 0.0741625651717186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16373020.0, "reward": 0.7059379816055298, "reward_std": 0.4659270942211151, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5368750095367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3687500059604645, "step": 1413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 408.4375, "completions/mean_terminated_length": 408.4375, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.2828, "grad_norm": 0.18495845794677734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16383523.0, "reward": 1.2467197179794312, "reward_std": 0.4777795970439911, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8059375286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.856249988079071, "step": 1414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 718.375, "completions/mean_terminated_length": 698.0000610351562, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.283, "grad_norm": 0.06605595350265503, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16399193.0, "reward": 0.6792947053909302, "reward_std": 0.34256288409233093, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35858750343322754, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6775000095367432, "step": 1415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 475.5625, "completions/mean_terminated_length": 475.5625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.2832, "grad_norm": 0.14304189383983612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16410898.0, "reward": 0.9794477820396423, "reward_std": 0.4792168140411377, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5213937759399414, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9206249713897705, "step": 1416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 316.0, "completions/max_terminated_length": 316.0, "completions/mean_length": 260.1875, "completions/mean_terminated_length": 260.1875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2834, "grad_norm": 0.00026582085411064327, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16418693.0, "reward": 1.4993197917938232, "reward_std": 0.0005911477492190897, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986374974250793, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 1417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 256.6875, "completions/mean_terminated_length": 256.6875, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.2836, "grad_norm": 0.18469609320163727, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16426512.0, "reward": 0.7323229312896729, "reward_std": 0.37666720151901245, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27714377641677856, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 1418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 364.8125, "completions/mean_terminated_length": 364.8125, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.2838, "grad_norm": 0.22649018466472626, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16436125.0, "reward": 1.1364822387695312, "reward_std": 0.48340344429016113, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6479625701904297, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 946.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 719.75, "completions/mean_terminated_length": 719.75, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.284, "grad_norm": 0.07610441744327545, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16451129.0, "reward": 0.8236509561538696, "reward_std": 0.4142455756664276, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39730000495910645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 1420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 316.0, "completions/max_terminated_length": 316.0, "completions/mean_length": 221.3125, "completions/mean_terminated_length": 221.3125, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.2842, "grad_norm": 0.263990193605423, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16458014.0, "reward": 0.7729604840278625, "reward_std": 0.4333842396736145, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2959187626838684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 1421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 456.9375, "completions/mean_terminated_length": 419.13336181640625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2844, "grad_norm": 0.12056101113557816, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16469597.0, "reward": 0.9141979217529297, "reward_std": 0.5145638585090637, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5158936977386475, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.815625011920929, "step": 1422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 565.625, "completions/mean_terminated_length": 565.625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.2846, "grad_norm": 0.14714652299880981, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16482839.0, "reward": 1.0737197399139404, "reward_std": 0.4985300302505493, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5849375128746033, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 425.4375, "completions/mean_terminated_length": 385.5333557128906, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.2848, "grad_norm": 0.09535340219736099, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16493662.0, "reward": 0.8095195293426514, "reward_std": 0.4193126857280731, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4940375089645386, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7049999833106995, "step": 1424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 533.0, "completions/max_terminated_length": 533.0, "completions/mean_length": 411.5625, "completions/mean_terminated_length": 411.5625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.285, "grad_norm": 0.09793207794427872, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16504775.0, "reward": 1.2876572608947754, "reward_std": 0.3894813060760498, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8253124952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9124999642372131, "step": 1425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 415.4375, "completions/mean_terminated_length": 415.4375, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.2852, "grad_norm": 0.10104028880596161, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16515262.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 460.0, "completions/mean_terminated_length": 460.0, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.2854, "grad_norm": 0.014821372926235199, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16526302.0, "reward": 0.5539075136184692, "reward_std": 0.08865498006343842, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10781249403953552, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 1427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 381.0, "completions/max_terminated_length": 381.0, "completions/mean_length": 300.25, "completions/mean_terminated_length": 300.25, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.2856, "grad_norm": 0.09498673677444458, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16534898.0, "reward": 0.5935759544372559, "reward_std": 0.24212723970413208, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1246500015258789, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 397.5, "completions/mean_terminated_length": 397.5, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.2858, "grad_norm": 0.1053125187754631, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16545834.0, "reward": 0.6573292016983032, "reward_std": 0.3294266164302826, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18965624272823334, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 1429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 398.0, "completions/mean_terminated_length": 398.0, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.286, "grad_norm": 0.06134290620684624, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16555962.0, "reward": 0.628594696521759, "reward_std": 0.24621354043483734, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19468750059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 1430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 434.5, "completions/mean_terminated_length": 434.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.2862, "grad_norm": 0.09615445882081985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16566450.0, "reward": 0.6583605408668518, "reward_std": 0.328623503446579, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19171875715255737, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 1431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 333.25, "completions/mean_terminated_length": 333.25, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.2864, "grad_norm": 0.15197636187076569, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16576390.0, "reward": 0.7738011479377747, "reward_std": 0.4333946108818054, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29760000109672546, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 770.125, "completions/mean_terminated_length": 753.2000122070312, "completions/min_length": 512.0, "completions/min_terminated_length": 512.0, "epoch": 0.2866, "grad_norm": 0.09014847874641418, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16592728.0, "reward": 0.839519739151001, "reward_std": 0.49550244212150574, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.42903751134872437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8362499475479126, "step": 1433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 432.0, "completions/max_terminated_length": 432.0, "completions/mean_length": 322.25, "completions/mean_terminated_length": 322.25, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.2868, "grad_norm": 0.21242061257362366, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16601516.0, "reward": 1.2565414905548096, "reward_std": 0.4337286055088043, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7630812525749207, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 480.875, "completions/mean_terminated_length": 480.875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.287, "grad_norm": 0.11491425335407257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16613770.0, "reward": 1.2816414833068848, "reward_std": 0.395169198513031, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8132812976837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 1435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 810.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 591.25, "completions/mean_terminated_length": 591.25, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.2872, "grad_norm": 0.09017694741487503, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16627070.0, "reward": 1.016951084136963, "reward_std": 0.41713041067123413, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6589000225067139, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7649999856948853, "step": 1436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 358.0, "completions/max_terminated_length": 358.0, "completions/mean_length": 288.5, "completions/mean_terminated_length": 288.5, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.2874, "grad_norm": 0.15236811339855194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16635606.0, "reward": 1.2561417818069458, "reward_std": 0.43500837683677673, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7622812390327454, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 1437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 697.5625, "completions/mean_terminated_length": 650.9285888671875, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.2876, "grad_norm": 0.02624776028096676, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16651455.0, "reward": 0.4731261432170868, "reward_std": 0.18971174955368042, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07125000655651093, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 1438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.2878, "grad_norm": 0.1540921926498413, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16662068.0, "reward": 1.3237512111663818, "reward_std": 0.3769854009151459, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8349999785423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 507.1875, "completions/mean_terminated_length": 472.7333679199219, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.288, "grad_norm": 0.07150636613368988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16674087.0, "reward": 0.9253910779953003, "reward_std": 0.32962489128112793, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.600781261920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 1440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 429.6875, "completions/mean_terminated_length": 429.6875, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.2882, "grad_norm": 0.14788343012332916, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16684754.0, "reward": 0.8402197360992432, "reward_std": 0.4589199125766754, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36793750524520874, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1023.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 529.0625, "completions/mean_terminated_length": 529.0625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.2884, "grad_norm": 0.026288727298378944, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16696787.0, "reward": 0.5689853429794312, "reward_std": 0.08342794328927994, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1379687488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 1442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 461.125, "completions/mean_terminated_length": 461.125, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.2886, "grad_norm": 0.1376226246356964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16708309.0, "reward": 0.8972511291503906, "reward_std": 0.4818030297756195, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4194999933242798, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 378.125, "completions/mean_terminated_length": 378.125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.2888, "grad_norm": 0.08004765212535858, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16718487.0, "reward": 1.3750543594360352, "reward_std": 0.339648962020874, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8751062154769897, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 1444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 256.9375, "completions/mean_terminated_length": 256.9375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.289, "grad_norm": 0.00023875113402027637, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16727318.0, "reward": 1.4991135597229004, "reward_std": 0.0005580336437560618, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982249736785889, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 1445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 321.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 261.3125, "completions/mean_terminated_length": 261.3125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.2892, "grad_norm": 0.2237124890089035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16735483.0, "reward": 1.1349074840545654, "reward_std": 0.48598596453666687, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6448125243186951, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 381.5625, "completions/mean_terminated_length": 381.5625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2894, "grad_norm": 0.10391344875097275, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16746148.0, "reward": 0.7232666015625, "reward_std": 0.37480849027633667, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2590312659740448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9231250286102295, "step": 1447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 510.375, "completions/mean_terminated_length": 510.375, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.2896, "grad_norm": 0.06876396387815475, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16758026.0, "reward": 0.6923447251319885, "reward_std": 0.22214002907276154, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.38468751311302185, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.675000011920929, "step": 1448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 355.6875, "completions/mean_terminated_length": 355.6875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.2898, "grad_norm": 0.07491478323936462, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16767285.0, "reward": 0.5959885716438293, "reward_std": 0.24148595333099365, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12947499752044678, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 1449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 780.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 623.0, "completions/mean_terminated_length": 623.0, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.29, "grad_norm": 0.11011629551649094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16781045.0, "reward": 1.0406259298324585, "reward_std": 0.47417208552360535, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.581250011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 1450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 566.375, "completions/mean_terminated_length": 566.375, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.2902, "grad_norm": 0.10616050660610199, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16794059.0, "reward": 1.0785073041915894, "reward_std": 0.4927663207054138, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5945124626159668, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 575.0, "completions/mean_terminated_length": 545.0667114257812, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.2904, "grad_norm": 0.02645489200949669, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16807435.0, "reward": 0.5124884247779846, "reward_std": 0.13794726133346558, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0874750018119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 1452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 358.125, "completions/mean_terminated_length": 313.73333740234375, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.2906, "grad_norm": 0.16682812571525574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16816813.0, "reward": 0.7246100902557373, "reward_std": 0.3413323760032654, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.57421875, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.390625, "step": 1453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 690.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 465.25, "completions/mean_terminated_length": 465.25, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.2908, "grad_norm": 0.044305261224508286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16829793.0, "reward": 0.9049222469329834, "reward_std": 0.15833231806755066, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8098437786102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3218750059604645, "step": 1454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 959.9375, "completions/mean_terminated_length": 910.1111450195312, "completions/min_length": 827.0, "completions/min_terminated_length": 827.0, "epoch": 0.291, "grad_norm": 0.06084664538502693, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16849664.0, "reward": 0.45609405636787415, "reward_std": 0.43148863315582275, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.34968751668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.3062500059604645, "step": 1455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 389.9375, "completions/mean_terminated_length": 389.9375, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.2912, "grad_norm": 0.13637307286262512, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16860143.0, "reward": 1.012866735458374, "reward_std": 0.5022289156913757, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5257312655448914, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 287.0625, "completions/mean_terminated_length": 287.0625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.2914, "grad_norm": 0.18101289868354797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16868640.0, "reward": 1.1703914403915405, "reward_std": 0.45027002692222595, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7157812714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 1457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 455.125, "completions/mean_terminated_length": 455.125, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.2916, "grad_norm": 0.1289621889591217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16880354.0, "reward": 1.1945606470108032, "reward_std": 0.467464417219162, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7016187310218811, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 1458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 470.8125, "completions/mean_terminated_length": 433.933349609375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2918, "grad_norm": 0.12506362795829773, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16891775.0, "reward": 1.3433539867401123, "reward_std": 0.4349413812160492, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8742063045501709, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 251.1875, "completions/mean_terminated_length": 251.1875, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.292, "grad_norm": 0.23330183327198029, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16899538.0, "reward": 0.9048447608947754, "reward_std": 0.475441575050354, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4346874952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 248.75, "completions/mean_terminated_length": 248.75, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.2922, "grad_norm": 0.06999924033880234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16907022.0, "reward": 1.437788724899292, "reward_std": 0.2474239617586136, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9380750060081482, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 1461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 415.4375, "completions/mean_terminated_length": 374.86669921875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.2924, "grad_norm": 0.160630002617836, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16917397.0, "reward": 1.2572664022445679, "reward_std": 0.46806520223617554, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8270312547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.828125, "step": 1462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 509.625, "completions/mean_terminated_length": 509.625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.2926, "grad_norm": 0.08549146354198456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16929775.0, "reward": 0.7267197370529175, "reward_std": 0.3834274113178253, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2659375071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 297.5625, "completions/mean_terminated_length": 297.5625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.2928, "grad_norm": 0.1769055426120758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16938136.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 243.25, "completions/mean_terminated_length": 243.25, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.293, "grad_norm": 0.16441120207309723, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16945500.0, "reward": 1.2573853731155396, "reward_std": 0.4327399730682373, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7647687196731567, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 563.0, "completions/mean_terminated_length": 563.0, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.2932, "grad_norm": 0.06683884561061859, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16958908.0, "reward": 0.6513264775276184, "reward_std": 0.33526796102523804, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17765000462532043, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 456.8125, "completions/mean_terminated_length": 456.8125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.2934, "grad_norm": 0.2045968472957611, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16970761.0, "reward": 0.840613603591919, "reward_std": 0.4587874710559845, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3687250018119812, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 441.5625, "completions/mean_terminated_length": 441.5625, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.2936, "grad_norm": 0.11094620823860168, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16981810.0, "reward": 1.2048447132110596, "reward_std": 0.4505309760570526, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7221875190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 653.625, "completions/mean_terminated_length": 600.7142944335938, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.2938, "grad_norm": 0.13581860065460205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16996124.0, "reward": 1.0395852327346802, "reward_std": 0.5818331241607666, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6416687965393066, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7881250381469727, "step": 1469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 999.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 727.0, "completions/mean_terminated_length": 727.0, "completions/min_length": 532.0, "completions/min_terminated_length": 532.0, "epoch": 0.294, "grad_norm": 0.042025044560432434, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17012268.0, "reward": 0.536716639995575, "reward_std": 0.18372897803783417, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13593125343322754, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8431249856948853, "step": 1470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 705.375, "completions/mean_terminated_length": 684.1333618164062, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.2942, "grad_norm": 0.046236760914325714, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17027666.0, "reward": 0.5356853604316711, "reward_std": 0.18527506291866302, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1338687539100647, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8381249904632568, "step": 1471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 663.75, "completions/mean_terminated_length": 639.7333374023438, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.2944, "grad_norm": 0.08070093393325806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17042190.0, "reward": 0.7304694652557373, "reward_std": 0.3392970561981201, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4609375, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.612500011920929, "step": 1472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 690.8125, "completions/mean_terminated_length": 668.6000366210938, "completions/min_length": 501.0, "completions/min_terminated_length": 501.0, "epoch": 0.2946, "grad_norm": 0.0543132983148098, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17057307.0, "reward": 1.2376289367675781, "reward_std": 0.44795238971710205, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8502562046051025, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.784375011920929, "step": 1473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 369.25, "completions/mean_terminated_length": 369.25, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.2948, "grad_norm": 0.1399454027414322, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17066703.0, "reward": 1.1366885900497437, "reward_std": 0.4835685193538666, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6483750343322754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 360.375, "completions/mean_terminated_length": 360.375, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.295, "grad_norm": 0.16290248930454254, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17076325.0, "reward": 0.7833575010299683, "reward_std": 0.4278368055820465, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31671249866485596, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 466.25, "completions/mean_terminated_length": 429.0666809082031, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.2952, "grad_norm": 0.13799597322940826, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17088489.0, "reward": 0.9301571846008301, "reward_std": 0.5350061655044556, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48531249165534973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 1476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 891.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 547.4375, "completions/mean_terminated_length": 547.4375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.2954, "grad_norm": 0.03452525660395622, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17103088.0, "reward": 0.5858606100082397, "reward_std": 0.15074540674686432, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1717187464237213, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 1477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 659.0625, "completions/mean_terminated_length": 537.4166870117188, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.2956, "grad_norm": 0.0965837761759758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17117281.0, "reward": 0.6251574754714966, "reward_std": 0.5295459628105164, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3128125071525574, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.637499988079071, "step": 1478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 357.5, "completions/mean_terminated_length": 357.5, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.2958, "grad_norm": 0.1811552792787552, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17126873.0, "reward": 1.3146073818206787, "reward_std": 0.39622873067855835, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8167125582695007, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 332.4375, "completions/mean_terminated_length": 332.4375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.296, "grad_norm": 0.16056498885154724, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17135952.0, "reward": 1.2616634368896484, "reward_std": 0.4250991940498352, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7733250260353088, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 1480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 318.9375, "completions/mean_terminated_length": 318.9375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.2962, "grad_norm": 0.1599758118391037, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17146575.0, "reward": 1.1351261138916016, "reward_std": 0.4861339032649994, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6452499628067017, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 1481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 518.25, "completions/mean_terminated_length": 518.25, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.2964, "grad_norm": 0.027614224702119827, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17158995.0, "reward": 0.5670262575149536, "reward_std": 0.12704409658908844, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13404999673366547, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 1482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 622.375, "completions/mean_terminated_length": 622.375, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.2966, "grad_norm": 0.12282592803239822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17173689.0, "reward": 1.3178510665893555, "reward_std": 0.38977837562561035, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.823199987411499, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 342.625, "completions/mean_terminated_length": 342.625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.2968, "grad_norm": 0.16729119420051575, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17183683.0, "reward": 1.3281102180480957, "reward_std": 0.3211985230445862, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.9062187671661377, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8293750286102295, "step": 1484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 463.9375, "completions/mean_terminated_length": 463.9375, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.297, "grad_norm": 0.023106422275304794, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17194754.0, "reward": 0.576482355594635, "reward_std": 0.0865822359919548, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15296249091625214, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 1485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 423.875, "completions/mean_terminated_length": 423.875, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.2972, "grad_norm": 0.1374068409204483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17205520.0, "reward": 1.035295009613037, "reward_std": 0.4929955005645752, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5705875158309937, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9162499904632568, "step": 1486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 508.625, "completions/mean_terminated_length": 508.625, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.2974, "grad_norm": 0.0616130456328392, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17217322.0, "reward": 1.4426040649414062, "reward_std": 0.22602808475494385, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9477062225341797, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 1487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 917.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 648.0625, "completions/mean_terminated_length": 648.0625, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.2976, "grad_norm": 0.08798666298389435, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17231643.0, "reward": 0.8257817029953003, "reward_std": 0.2852649390697479, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7140624523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 1488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 580.625, "completions/mean_terminated_length": 580.625, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.2978, "grad_norm": 0.18326714634895325, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17246021.0, "reward": 1.2203917503356934, "reward_std": 0.5040379762649536, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7532812356948853, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.878125011920929, "step": 1489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 358.0, "completions/max_terminated_length": 358.0, "completions/mean_length": 290.375, "completions/mean_terminated_length": 290.375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.298, "grad_norm": 0.23399119079113007, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17254667.0, "reward": 0.7862509489059448, "reward_std": 0.4248528778553009, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 769.8125, "completions/mean_terminated_length": 654.2727661132812, "completions/min_length": 487.0, "completions/min_terminated_length": 487.0, "epoch": 0.2982, "grad_norm": 0.13584236800670624, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17271224.0, "reward": 0.9741132259368896, "reward_std": 0.6949409246444702, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6357250213623047, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5625, "step": 1491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 461.5625, "completions/mean_terminated_length": 461.5625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.2984, "grad_norm": 0.0795055627822876, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17282577.0, "reward": 1.3409384489059448, "reward_std": 0.3342260718345642, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8693749904632568, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 1492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 598.6875, "completions/mean_terminated_length": 598.6875, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.2986, "grad_norm": 0.0042205266654491425, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17295916.0, "reward": 0.5390385985374451, "reward_std": 0.018110457807779312, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07807499170303345, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 1493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 428.5625, "completions/mean_terminated_length": 428.5625, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.2988, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17307061.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 455.125, "completions/mean_terminated_length": 455.125, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.299, "grad_norm": 0.07478608191013336, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17318247.0, "reward": 1.3800665140151978, "reward_std": 0.3245125114917755, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 353.9375, "completions/mean_terminated_length": 353.9375, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.2992, "grad_norm": 0.14277753233909607, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17328422.0, "reward": 0.9627853631973267, "reward_std": 0.4886854588985443, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48806875944137573, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 1496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 793.1875, "completions/mean_terminated_length": 777.800048828125, "completions/min_length": 597.0, "completions/min_terminated_length": 597.0, "epoch": 0.2994, "grad_norm": 0.04928687959909439, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17345481.0, "reward": 0.737722635269165, "reward_std": 0.32295888662338257, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4754437506198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6331250071525574, "step": 1497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 468.3125, "completions/mean_terminated_length": 431.2666931152344, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.2996, "grad_norm": 0.15375475585460663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17357118.0, "reward": 0.9329695701599121, "reward_std": 0.532840371131897, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4909375011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 1498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 241.75, "completions/mean_terminated_length": 241.75, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.2998, "grad_norm": 0.00029473358881659806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17364986.0, "reward": 1.4991729259490967, "reward_std": 0.0005645170458592474, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983437657356262, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 353.0, "completions/mean_terminated_length": 353.0, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.3, "grad_norm": 0.10045517981052399, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17374506.0, "reward": 0.6589823365211487, "reward_std": 0.3283131420612335, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19296249747276306, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 329.0, "completions/max_terminated_length": 329.0, "completions/mean_length": 275.0, "completions/mean_terminated_length": 275.0, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.3002, "grad_norm": 0.0963197723031044, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17382362.0, "reward": 0.5808918476104736, "reward_std": 0.2457764744758606, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09928125143051147, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 1501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 377.5625, "completions/mean_terminated_length": 377.5625, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.3004, "grad_norm": 0.18060429394245148, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17392499.0, "reward": 1.0102667808532715, "reward_std": 0.5055762529373169, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5205312967300415, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 1502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 305.8125, "completions/mean_terminated_length": 305.8125, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.3006, "grad_norm": 0.00014160395949147642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17401696.0, "reward": 1.4989008903503418, "reward_std": 0.0004099227662663907, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977999925613403, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 404.5, "completions/mean_terminated_length": 404.5, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.3008, "grad_norm": 0.02888736128807068, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17412056.0, "reward": 0.5923447608947754, "reward_std": 0.10214431583881378, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18468749523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 1504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 334.0, "completions/mean_terminated_length": 334.0, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.301, "grad_norm": 0.14726586639881134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17421352.0, "reward": 1.4368386268615723, "reward_std": 0.24982374906539917, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9361749887466431, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 766.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 477.5, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.3012, "grad_norm": 0.07135120779275894, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17433088.0, "reward": 1.305778980255127, "reward_std": 0.2973538041114807, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.9240562319755554, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7806249856948853, "step": 1506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 568.1875, "completions/mean_terminated_length": 503.0714416503906, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.3014, "grad_norm": 0.15329448878765106, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17446243.0, "reward": 0.6757040023803711, "reward_std": 0.459207683801651, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2889062464237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.809374988079071, "step": 1507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 240.0, "completions/max_terminated_length": 240.0, "completions/mean_length": 177.5625, "completions/mean_terminated_length": 177.5625, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 0.3016, "grad_norm": 0.3028424084186554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17452508.0, "reward": 1.3166855573654175, "reward_std": 0.39392077922821045, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.820868730545044, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.986875057220459, "step": 1508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 432.0625, "completions/mean_terminated_length": 432.0625, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.3018, "grad_norm": 0.12345315515995026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17463485.0, "reward": 1.3723664283752441, "reward_std": 0.2272791564464569, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.9947312474250793, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7593750357627869, "step": 1509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 367.125, "completions/mean_terminated_length": 367.125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.302, "grad_norm": 0.16079780459403992, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17473055.0, "reward": 0.9050759673118591, "reward_std": 0.4751001000404358, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43514999747276306, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 427.9375, "completions/mean_terminated_length": 388.20001220703125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.3022, "grad_norm": 0.04770277813076973, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17483582.0, "reward": 0.5543668270111084, "reward_std": 0.2848561108112335, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10873124748468399, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 1511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 523.6875, "completions/mean_terminated_length": 490.3333740234375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.3024, "grad_norm": 0.14452330768108368, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17496505.0, "reward": 0.8294537663459778, "reward_std": 0.39597824215888977, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5339062213897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 1512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 590.4375, "completions/mean_terminated_length": 561.5333862304688, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.3026, "grad_norm": 0.09358642995357513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17509808.0, "reward": 1.2520947456359863, "reward_std": 0.47795370221138, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8166875243186951, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8512499928474426, "step": 1513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 288.0, "completions/max_terminated_length": 288.0, "completions/mean_length": 260.5, "completions/mean_terminated_length": 260.5, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.3028, "grad_norm": 0.37607595324516296, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17517960.0, "reward": 0.8385759592056274, "reward_std": 0.4604262113571167, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3646499812602997, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 827.9375, "completions/mean_terminated_length": 814.86669921875, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.303, "grad_norm": 0.06890934705734253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17536855.0, "reward": 0.522572934627533, "reward_std": 0.25931739807128906, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1701437532901764, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7862499952316284, "step": 1515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 247.0, "completions/max_terminated_length": 247.0, "completions/mean_length": 204.75, "completions/mean_terminated_length": 204.75, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.3032, "grad_norm": 0.0002741506032180041, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17544115.0, "reward": 1.4995760917663574, "reward_std": 0.0005763607914559543, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991500377655029, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 1516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 607.0, "completions/mean_terminated_length": 510.7692565917969, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.3034, "grad_norm": 0.12141458690166473, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17557683.0, "reward": 0.7001570463180542, "reward_std": 0.5307035446166992, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33781248331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.75, "step": 1517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 401.1875, "completions/mean_terminated_length": 401.1875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.3036, "grad_norm": 0.18429185450077057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17568294.0, "reward": 0.8456259369850159, "reward_std": 0.4547778367996216, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3787499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 524.3125, "completions/mean_terminated_length": 524.3125, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.3038, "grad_norm": 0.11625174432992935, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17580891.0, "reward": 1.0430760383605957, "reward_std": 0.5492984056472778, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5861499905586243, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.96875, "step": 1519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 430.8125, "completions/mean_terminated_length": 430.8125, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.304, "grad_norm": 0.1151440441608429, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17592008.0, "reward": 0.6129512786865234, "reward_std": 0.2651989161968231, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16339999437332153, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9225000143051147, "step": 1520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 345.4375, "completions/mean_terminated_length": 345.4375, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.3042, "grad_norm": 0.1567346751689911, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17601471.0, "reward": 1.2614666223526, "reward_std": 0.4249817132949829, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7729312777519226, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 278.25, "completions/mean_terminated_length": 278.25, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.3044, "grad_norm": 0.1585957109928131, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17610979.0, "reward": 0.7069450616836548, "reward_std": 0.3938397765159607, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22638750076293945, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "step": 1522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 569.1875, "completions/mean_terminated_length": 464.23077392578125, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.3046, "grad_norm": 0.16169211268424988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17623974.0, "reward": 0.887344479560852, "reward_std": 0.6065974831581116, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5246875286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.731249988079071, "step": 1523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 417.8125, "completions/mean_terminated_length": 417.8125, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.3048, "grad_norm": 0.18297941982746124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17634787.0, "reward": 0.7996165752410889, "reward_std": 0.5048542022705078, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3492310047149658, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9748749732971191, "step": 1524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 501.25, "completions/mean_terminated_length": 501.25, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.305, "grad_norm": 0.0059372312389314175, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17646823.0, "reward": 0.5396105051040649, "reward_std": 0.019651303067803383, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 534.25, "completions/mean_terminated_length": 501.60003662109375, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.3052, "grad_norm": 0.14207962155342102, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17659003.0, "reward": 0.7179695963859558, "reward_std": 0.4254806637763977, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3109374940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 1526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 839.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 608.5, "completions/mean_terminated_length": 608.5, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.3054, "grad_norm": 0.08254413306713104, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17672723.0, "reward": 0.8425791263580322, "reward_std": 0.4570569694042206, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3726562559604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 542.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 370.3125, "completions/mean_terminated_length": 370.3125, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.3056, "grad_norm": 0.04349033161997795, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17682632.0, "reward": 1.4510140419006348, "reward_std": 0.19527070224285126, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9645249843597412, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 1528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 435.5, "completions/mean_terminated_length": 435.5, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.3058, "grad_norm": 0.11463107913732529, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17693696.0, "reward": 1.2614822387695312, "reward_std": 0.4249911606311798, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7729625105857849, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 449.25, "completions/mean_terminated_length": 449.25, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.306, "grad_norm": 0.16054309904575348, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17705236.0, "reward": 0.8717852830886841, "reward_std": 0.4505726099014282, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4310687482357025, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 1530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 312.25, "completions/mean_terminated_length": 312.25, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.3062, "grad_norm": 0.0505710244178772, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17713896.0, "reward": 1.4596071243286133, "reward_std": 0.1558983325958252, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9817125201225281, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 1531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 465.5625, "completions/mean_terminated_length": 465.5625, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.3064, "grad_norm": 0.10091853141784668, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17725409.0, "reward": 0.5317978858947754, "reward_std": 0.16438636183738708, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12609374523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 1532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 499.0, "completions/max_terminated_length": 499.0, "completions/mean_length": 461.625, "completions/mean_terminated_length": 461.625, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.3066, "grad_norm": 0.00016693676298018545, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17736603.0, "reward": 1.4995605945587158, "reward_std": 0.0005680285976268351, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991187453269958, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 1533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 466.0, "completions/max_terminated_length": 466.0, "completions/mean_length": 252.8125, "completions/mean_terminated_length": 252.8125, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.3068, "grad_norm": 0.12979966402053833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17744440.0, "reward": 1.3184854984283447, "reward_std": 0.3900929391384125, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.824468731880188, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 1534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 267.5, "completions/mean_terminated_length": 267.5, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.307, "grad_norm": 0.1219828650355339, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17752688.0, "reward": 0.6588261127471924, "reward_std": 0.3288270831108093, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1926499903202057, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 958.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 588.3125, "completions/mean_terminated_length": 588.3125, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.3072, "grad_norm": 0.1070164144039154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17765797.0, "reward": 1.0504666566848755, "reward_std": 0.3997877240180969, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7259312272071838, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6931250095367432, "step": 1536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 298.0625, "completions/mean_terminated_length": 298.0625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.3074, "grad_norm": 0.16877920925617218, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17775238.0, "reward": 1.0773324966430664, "reward_std": 0.49436885118484497, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5921624898910522, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 479.3125, "completions/mean_terminated_length": 401.5000305175781, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.3076, "grad_norm": 0.1441367119550705, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17787131.0, "reward": 1.3114070892333984, "reward_std": 0.5119208693504333, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8728125095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 1538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 644.5, "completions/mean_terminated_length": 619.2000122070312, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.3078, "grad_norm": 0.07475893944501877, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17801731.0, "reward": 0.6077351570129395, "reward_std": 0.23916712403297424, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.27796873450279236, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.715624988079071, "step": 1539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 413.8125, "completions/mean_terminated_length": 413.8125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.308, "grad_norm": 0.13539397716522217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17812176.0, "reward": 1.0133479833602905, "reward_std": 0.5013870596885681, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5266937613487244, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 550.0625, "completions/mean_terminated_length": 550.0625, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.3082, "grad_norm": 0.0837898924946785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17825137.0, "reward": 1.3802292346954346, "reward_std": 0.3245762288570404, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854562640190125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 534.0625, "completions/mean_terminated_length": 501.4000244140625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.3084, "grad_norm": 0.11205445975065231, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17837442.0, "reward": 0.6684389710426331, "reward_std": 0.3920517861843109, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27437499165534973, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.800000011920929, "step": 1542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 505.0, "completions/mean_terminated_length": 470.4000244140625, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.3086, "grad_norm": 0.17361809313297272, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17849650.0, "reward": 0.9925791621208191, "reward_std": 0.540287971496582, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.547656238079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 1543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 710.5, "completions/mean_terminated_length": 689.6000366210938, "completions/min_length": 552.0, "completions/min_terminated_length": 552.0, "epoch": 0.3088, "grad_norm": 0.07622627913951874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17865866.0, "reward": 1.254922866821289, "reward_std": 0.4728485345840454, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8223437070846558, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 1544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 896.375, "completions/mean_terminated_length": 797.1111450195312, "completions/min_length": 676.0, "completions/min_terminated_length": 676.0, "epoch": 0.309, "grad_norm": 0.04842127114534378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17884848.0, "reward": 0.30366623401641846, "reward_std": 0.2774112820625305, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04483124986290932, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5393749475479126, "step": 1545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 580.25, "completions/mean_terminated_length": 580.25, "completions/min_length": 454.0, "completions/min_terminated_length": 454.0, "epoch": 0.3092, "grad_norm": 0.045180048793554306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17897956.0, "reward": 0.6020323634147644, "reward_std": 0.23969846963882446, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14156249165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 445.0, "completions/max_terminated_length": 445.0, "completions/mean_length": 387.0, "completions/mean_terminated_length": 387.0, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.3094, "grad_norm": 0.1055721566081047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17908148.0, "reward": 1.439507246017456, "reward_std": 0.2375352531671524, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 931.25, "completions/mean_terminated_length": 889.0909423828125, "completions/min_length": 769.0, "completions/min_terminated_length": 769.0, "epoch": 0.3096, "grad_norm": 0.09346958994865417, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17927352.0, "reward": 0.8457162380218506, "reward_std": 0.674646258354187, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5664312839508057, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5556250214576721, "step": 1548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 323.5625, "completions/mean_terminated_length": 323.5625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.3098, "grad_norm": 0.26130351424217224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17936929.0, "reward": 0.9876446723937988, "reward_std": 0.5440539717674255, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5377874970436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 376.25, "completions/mean_terminated_length": 376.25, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.31, "grad_norm": 0.15920591354370117, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17946533.0, "reward": 0.9504637718200684, "reward_std": 0.49986985325813293, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46342504024505615, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 269.5625, "completions/mean_terminated_length": 269.5625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.3102, "grad_norm": 0.20746631920337677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17954782.0, "reward": 1.1995983123779297, "reward_std": 0.4601186513900757, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7116937637329102, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 1551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 276.5625, "completions/mean_terminated_length": 276.5625, "completions/min_length": 119.0, "completions/min_terminated_length": 119.0, "epoch": 0.3104, "grad_norm": 0.29664507508277893, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17962951.0, "reward": 1.0163886547088623, "reward_std": 0.4995122253894806, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5327750444412231, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 1552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 643.1875, "completions/mean_terminated_length": 643.1875, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.3106, "grad_norm": 0.060220904648303986, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17977530.0, "reward": 0.6103105545043945, "reward_std": 0.26399365067481995, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15811875462532043, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9318749904632568, "step": 1553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 542.4375, "completions/mean_terminated_length": 431.3077087402344, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.3108, "grad_norm": 0.1309741884469986, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17990273.0, "reward": 0.693816602230072, "reward_std": 0.509031355381012, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.38763123750686646, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6331250071525574, "step": 1554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 578.4375, "completions/mean_terminated_length": 578.4375, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.311, "grad_norm": 0.11060597002506256, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18003608.0, "reward": 0.9548853635787964, "reward_std": 0.49606192111968994, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4722687602043152, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 400.5625, "completions/mean_terminated_length": 400.5625, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.3112, "grad_norm": 0.09583805501461029, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18014433.0, "reward": 1.3189728260040283, "reward_std": 0.38697323203086853, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8254437446594238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 459.0625, "completions/mean_terminated_length": 459.0625, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.3114, "grad_norm": 0.13708670437335968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18025650.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 360.25, "completions/mean_terminated_length": 360.25, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.3116, "grad_norm": 0.2189004123210907, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18035142.0, "reward": 1.3151917457580566, "reward_std": 0.3965188264846802, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8178812265396118, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 1558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 804.9375, "completions/mean_terminated_length": 790.3333740234375, "completions/min_length": 636.0, "completions/min_terminated_length": 636.0, "epoch": 0.3118, "grad_norm": 0.031934287399053574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18052725.0, "reward": 0.8440629243850708, "reward_std": 0.25575605034828186, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7506250143051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 1559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 353.25, "completions/mean_terminated_length": 353.25, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.312, "grad_norm": 0.13913197815418243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18062601.0, "reward": 1.378638505935669, "reward_std": 0.32964828610420227, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8822749853134155, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 611.1875, "completions/mean_terminated_length": 611.1875, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.3122, "grad_norm": 0.08237355202436447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18076156.0, "reward": 1.1599979400634766, "reward_std": 0.45902135968208313, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6949937343597412, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 1561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 542.375, "completions/mean_terminated_length": 510.2666931152344, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.3124, "grad_norm": 0.10698580741882324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18089570.0, "reward": 0.8476539254188538, "reward_std": 0.46909403800964355, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.44530627131462097, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7556250095367432, "step": 1562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 269.25, "completions/mean_terminated_length": 269.25, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.3126, "grad_norm": 0.17110206186771393, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18097926.0, "reward": 1.0172353982925415, "reward_std": 0.4977933168411255, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5344687700271606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 1563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 619.1875, "completions/mean_terminated_length": 619.1875, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.3128, "grad_norm": 0.08788556605577469, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18112905.0, "reward": 0.7491415143013, "reward_std": 0.3712713420391083, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31078124046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 1564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 589.8125, "completions/mean_terminated_length": 589.8125, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.313, "grad_norm": 0.060495976358652115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18126358.0, "reward": 0.7109726071357727, "reward_std": 0.1970689594745636, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.421943724155426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6668750047683716, "step": 1565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 509.0, "completions/max_terminated_length": 509.0, "completions/mean_length": 434.9375, "completions/mean_terminated_length": 434.9375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.3132, "grad_norm": 0.07019384950399399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18137557.0, "reward": 0.6026977896690369, "reward_std": 0.23922792077064514, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428937464952469, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 345.0, "completions/max_terminated_length": 345.0, "completions/mean_length": 279.625, "completions/mean_terminated_length": 279.625, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.3134, "grad_norm": 0.007676380220800638, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18145647.0, "reward": 0.526963472366333, "reward_std": 0.01786973513662815, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05392500013113022, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 1567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 482.8125, "completions/mean_terminated_length": 482.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3136, "grad_norm": 0.17753247916698456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18157548.0, "reward": 0.9458571076393127, "reward_std": 0.4551238417625427, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5167125463485718, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 1568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 562.875, "completions/mean_terminated_length": 562.875, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.3138, "grad_norm": 0.1304672211408615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18170522.0, "reward": 0.8985384702682495, "reward_std": 0.4807782471179962, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4220750033855438, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 1569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 406.3125, "completions/mean_terminated_length": 406.3125, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.314, "grad_norm": 0.20499025285243988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18181407.0, "reward": 1.083482265472412, "reward_std": 0.48705488443374634, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6044625043869019, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 881.1875, "completions/mean_terminated_length": 833.5833740234375, "completions/min_length": 741.0, "completions/min_terminated_length": 741.0, "epoch": 0.3142, "grad_norm": 0.06581383943557739, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18199362.0, "reward": 0.5565476417541504, "reward_std": 0.4614275395870209, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23809373378753662, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6731250286102295, "step": 1571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 261.6875, "completions/mean_terminated_length": 261.6875, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.3144, "grad_norm": 0.13539506494998932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18207117.0, "reward": 1.3210698366165161, "reward_std": 0.38317787647247314, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8296375274658203, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 699.5, "completions/mean_terminated_length": 677.86669921875, "completions/min_length": 443.0, "completions/min_terminated_length": 443.0, "epoch": 0.3146, "grad_norm": 0.04749650880694389, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18222453.0, "reward": 0.6778007745742798, "reward_std": 0.3284592628479004, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2930999994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 1573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 283.5625, "completions/mean_terminated_length": 283.5625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.3148, "grad_norm": 0.00026299094315618277, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18231102.0, "reward": 1.4990415573120117, "reward_std": 0.0005210248637013137, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980812668800354, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 545.6875, "completions/mean_terminated_length": 545.6875, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.315, "grad_norm": 0.09773477911949158, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18243945.0, "reward": 1.198960304260254, "reward_std": 0.3911435604095459, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.772918701171875, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8068749904632568, "step": 1575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 592.25, "completions/mean_terminated_length": 592.25, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.3152, "grad_norm": 0.09245600551366806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18257357.0, "reward": 0.829688310623169, "reward_std": 0.3514551520347595, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.47187501192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 1576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 397.625, "completions/mean_terminated_length": 397.625, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.3154, "grad_norm": 0.14908692240715027, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18267703.0, "reward": 1.2224228382110596, "reward_std": 0.4302985370159149, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7573437690734863, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281249642372131, "step": 1577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 612.6875, "completions/mean_terminated_length": 585.2667236328125, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.3156, "grad_norm": 0.09649144858121872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18281170.0, "reward": 0.9803136587142944, "reward_std": 0.5506551265716553, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5231249928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 604.0625, "completions/mean_terminated_length": 604.0625, "completions/min_length": 450.0, "completions/min_terminated_length": 450.0, "epoch": 0.3158, "grad_norm": 0.11711914837360382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18294835.0, "reward": 1.022500991821289, "reward_std": 0.44214901328086853, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6074999570846558, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8125, "step": 1579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 903.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 744.9375, "completions/mean_terminated_length": 744.9375, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.316, "grad_norm": 0.03210991993546486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18311170.0, "reward": 0.6839852333068848, "reward_std": 0.12447965145111084, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.36796873807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.778124988079071, "step": 1580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 461.5, "completions/mean_terminated_length": 461.5, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.3162, "grad_norm": 0.07357719540596008, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18321994.0, "reward": 0.7465632557868958, "reward_std": 0.19278906285762787, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49312499165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6437499523162842, "step": 1581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 354.5, "completions/mean_terminated_length": 354.5, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.3164, "grad_norm": 0.20039556920528412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18332162.0, "reward": 1.077354073524475, "reward_std": 0.49418020248413086, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5922062397003174, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 377.125, "completions/mean_terminated_length": 377.125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.3166, "grad_norm": 0.09977979958057404, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18341924.0, "reward": 1.378147840499878, "reward_std": 0.33141258358955383, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.881293773651123, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 1583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 293.125, "completions/mean_terminated_length": 293.125, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.3168, "grad_norm": 0.3387174904346466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18350198.0, "reward": 1.0831259489059448, "reward_std": 0.48673015832901, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6037499904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 623.6875, "completions/mean_terminated_length": 597.0000610351562, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.317, "grad_norm": 0.04978477954864502, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18363889.0, "reward": 0.5735163688659668, "reward_std": 0.27128246426582336, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2720312476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7437500357627869, "step": 1585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 902.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 602.3125, "completions/mean_terminated_length": 602.3125, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.3172, "grad_norm": 0.03341512754559517, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18377654.0, "reward": 0.5774102210998535, "reward_std": 0.07844816893339157, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15481874346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143749475479126, "step": 1586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 551.9375, "completions/mean_terminated_length": 551.9375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.3174, "grad_norm": 0.10796110332012177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18390405.0, "reward": 0.7163915634155273, "reward_std": 0.42215248942375183, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.30778124928474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8606250286102295, "step": 1587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 436.6875, "completions/mean_terminated_length": 436.6875, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.3176, "grad_norm": 0.06077314540743828, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18401616.0, "reward": 1.4399573802947998, "reward_std": 0.23765574395656586, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424124956130981, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9699999690055847, "step": 1588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 817.75, "completions/mean_terminated_length": 788.2857666015625, "completions/min_length": 696.0, "completions/min_terminated_length": 696.0, "epoch": 0.3178, "grad_norm": 0.04250278323888779, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18418908.0, "reward": 0.5646884441375732, "reward_std": 0.33731234073638916, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19187501072883606, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 1589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 483.1875, "completions/mean_terminated_length": 483.1875, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.318, "grad_norm": 0.11294489353895187, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18430255.0, "reward": 1.2017977237701416, "reward_std": 0.4017846882343292, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7785937786102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.846875011920929, "step": 1590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 570.25, "completions/mean_terminated_length": 570.25, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.3182, "grad_norm": 0.07159840315580368, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18443347.0, "reward": 0.5735949873924255, "reward_std": 0.22091110050678253, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2096874862909317, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8406250476837158, "step": 1591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 691.25, "completions/mean_terminated_length": 669.0667114257812, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.3184, "grad_norm": 0.036814890801906586, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18458727.0, "reward": 0.5392978191375732, "reward_std": 0.182795450091362, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14109376072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8406249284744263, "step": 1592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 370.625, "completions/mean_terminated_length": 370.625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.3186, "grad_norm": 0.16403760015964508, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18468449.0, "reward": 0.9897664785385132, "reward_std": 0.474033921957016, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5420312881469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 1593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 453.0, "completions/mean_terminated_length": 453.0, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.3188, "grad_norm": 0.15241943299770355, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18479761.0, "reward": 0.9927290678024292, "reward_std": 0.4741184115409851, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5479562282562256, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 341.75, "completions/mean_terminated_length": 341.75, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.319, "grad_norm": 0.19707755744457245, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18489597.0, "reward": 1.2020196914672852, "reward_std": 0.45487818121910095, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7165374755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 288.0, "completions/mean_terminated_length": 288.0, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.3192, "grad_norm": 0.28732743859291077, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18497949.0, "reward": 0.9290605783462524, "reward_std": 0.4657750427722931, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4831187427043915, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 1596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 738.75, "completions/mean_terminated_length": 738.75, "completions/min_length": 515.0, "completions/min_terminated_length": 515.0, "epoch": 0.3194, "grad_norm": 0.031775474548339844, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18513609.0, "reward": 0.6020323634147644, "reward_std": 0.23969846963882446, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14156249165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 718.75, "completions/mean_terminated_length": 617.0, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.3196, "grad_norm": 0.06262877583503723, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18530021.0, "reward": 0.4518634080886841, "reward_std": 0.29645130038261414, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15372499823570251, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6549999713897705, "step": 1598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 565.625, "completions/mean_terminated_length": 565.625, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.3198, "grad_norm": 0.09948346018791199, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18542879.0, "reward": 1.3760137557983398, "reward_std": 0.3381151258945465, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8770250082015991, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9837499856948853, "step": 1599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 486.625, "completions/mean_terminated_length": 486.625, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.32, "grad_norm": 0.09235141426324844, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18554329.0, "reward": 0.7807791233062744, "reward_std": 0.42836809158325195, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31155624985694885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 579.625, "completions/mean_terminated_length": 579.625, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.3202, "grad_norm": 0.05087165907025337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18570083.0, "reward": 0.6466414928436279, "reward_std": 0.25524070858955383, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23078122735023499, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 1601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 498.0, "completions/mean_terminated_length": 498.0, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.3204, "grad_norm": 0.00295550306327641, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18581923.0, "reward": 0.5445165634155273, "reward_std": 0.011709419079124928, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 385.4375, "completions/mean_terminated_length": 385.4375, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.3206, "grad_norm": 0.10559413582086563, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18592794.0, "reward": 0.613632321357727, "reward_std": 0.2474229484796524, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1647624969482422, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 1603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 402.1875, "completions/mean_terminated_length": 402.1875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.3208, "grad_norm": 0.13533751666545868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18604237.0, "reward": 1.3346853256225586, "reward_std": 0.35419175028800964, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8568687438964844, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 1604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 347.0625, "completions/mean_terminated_length": 347.0625, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.321, "grad_norm": 0.13681359589099884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18614510.0, "reward": 0.7231072187423706, "reward_std": 0.38532117009162903, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.258712500333786, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 273.5, "completions/mean_terminated_length": 273.5, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.3212, "grad_norm": 0.18497073650360107, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18622966.0, "reward": 1.1995258331298828, "reward_std": 0.45863255858421326, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7115499973297119, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 286.125, "completions/mean_terminated_length": 286.125, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.3214, "grad_norm": 0.15118853747844696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18631320.0, "reward": 0.839041531085968, "reward_std": 0.45973411202430725, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3655812442302704, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 446.8125, "completions/mean_terminated_length": 364.3571472167969, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.3216, "grad_norm": 0.18351414799690247, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18642293.0, "reward": 1.0428884029388428, "reward_std": 0.5788663029670715, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6482750177383423, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7774999737739563, "step": 1608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 455.3125, "completions/mean_terminated_length": 455.3125, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.3218, "grad_norm": 0.1306600570678711, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18653178.0, "reward": 1.0320322513580322, "reward_std": 0.4468013346195221, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6265624761581421, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 1609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 475.0, "completions/max_terminated_length": 475.0, "completions/mean_length": 384.0, "completions/mean_terminated_length": 384.0, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.322, "grad_norm": 0.11147811263799667, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18666506.0, "reward": 0.64695143699646, "reward_std": 0.33378908038139343, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1688999980688095, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 1610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 657.25, "completions/mean_terminated_length": 604.857177734375, "completions/min_length": 460.0, "completions/min_terminated_length": 460.0, "epoch": 0.3222, "grad_norm": 0.08425933122634888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18681118.0, "reward": 0.6081916689872742, "reward_std": 0.40576228499412537, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2163812518119812, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8143749833106995, "step": 1611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 470.1875, "completions/mean_terminated_length": 470.1875, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.3224, "grad_norm": 0.11514105647802353, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18692881.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 435.75, "completions/mean_terminated_length": 435.75, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.3226, "grad_norm": 0.010100923478603363, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18703933.0, "reward": 0.9836723804473877, "reward_std": 0.031896643340587616, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9673437476158142, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.11562500894069672, "step": 1613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 546.125, "completions/mean_terminated_length": 546.125, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.3228, "grad_norm": 0.11349047720432281, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18716511.0, "reward": 0.5808353424072266, "reward_std": 0.24470022320747375, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22416874766349792, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7493749856948853, "step": 1614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 613.4375, "completions/mean_terminated_length": 613.4375, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.323, "grad_norm": 0.08988961577415466, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18730806.0, "reward": 0.952345073223114, "reward_std": 0.49836739897727966, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4671874940395355, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 1615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 507.3125, "completions/mean_terminated_length": 472.86669921875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.3232, "grad_norm": 0.10265721380710602, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18743179.0, "reward": 0.6846101880073547, "reward_std": 0.3367193341255188, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.36921876668930054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6781249642372131, "step": 1616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 350.6875, "completions/mean_terminated_length": 350.6875, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.3234, "grad_norm": 0.17566391825675964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18752342.0, "reward": 1.3187103271484375, "reward_std": 0.388510137796402, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8249187469482422, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 406.8125, "completions/mean_terminated_length": 406.8125, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.3236, "grad_norm": 0.1843404471874237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18762707.0, "reward": 1.1052979230880737, "reward_std": 0.540177047252655, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6480937600135803, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 373.625, "completions/mean_terminated_length": 373.625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.3238, "grad_norm": 0.15516941249370575, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18772045.0, "reward": 1.3207571506500244, "reward_std": 0.3830225169658661, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290125131607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 315.8125, "completions/mean_terminated_length": 315.8125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.324, "grad_norm": 0.124838687479496, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18780570.0, "reward": 0.6443700790405273, "reward_std": 0.3342634439468384, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16373750567436218, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 381.6875, "completions/mean_terminated_length": 338.86669921875, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.3242, "grad_norm": 0.16483598947525024, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18791381.0, "reward": 1.2254573106765747, "reward_std": 0.5042865872383118, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7634124755859375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 1621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 939.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 630.875, "completions/mean_terminated_length": 630.875, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.3244, "grad_norm": 0.09441335499286652, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18805715.0, "reward": 0.6268761157989502, "reward_std": 0.25962933897972107, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19124998152256012, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 1622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 493.1875, "completions/mean_terminated_length": 457.8000183105469, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.3246, "grad_norm": 0.20974338054656982, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18817270.0, "reward": 0.7993759512901306, "reward_std": 0.5339888334274292, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4112499952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.856249988079071, "step": 1623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 504.5625, "completions/mean_terminated_length": 504.5625, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.3248, "grad_norm": 0.07501524686813354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18831039.0, "reward": 0.7185478210449219, "reward_std": 0.38774293661117554, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24959373474121094, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 396.75, "completions/mean_terminated_length": 396.75, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.325, "grad_norm": 0.09709315747022629, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18841259.0, "reward": 1.3784135580062866, "reward_std": 0.3295603096485138, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8818249702453613, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 636.9375, "completions/mean_terminated_length": 581.6428833007812, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.3252, "grad_norm": 0.10994900017976761, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18855290.0, "reward": 0.833585262298584, "reward_std": 0.5614497065544128, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4171687364578247, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8418750166893005, "step": 1626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 314.125, "completions/mean_terminated_length": 314.125, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.3254, "grad_norm": 0.22453798353672028, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18864252.0, "reward": 1.009338617324829, "reward_std": 0.5059477686882019, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5186750292778015, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 527.1875, "completions/mean_terminated_length": 494.0666809082031, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.3256, "grad_norm": 0.1404348611831665, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18876991.0, "reward": 0.8113290071487427, "reward_std": 0.4973577857017517, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3726562559604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 1628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 626.1875, "completions/mean_terminated_length": 626.1875, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.3258, "grad_norm": 0.10510233789682388, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18890658.0, "reward": 1.0094012022018433, "reward_std": 0.5062158107757568, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5188000202178955, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 1629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 324.1875, "completions/mean_terminated_length": 324.1875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.326, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18899701.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 868.3125, "completions/mean_terminated_length": 747.2222290039062, "completions/min_length": 653.0, "completions/min_terminated_length": 653.0, "epoch": 0.3262, "grad_norm": 0.15151837468147278, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18917818.0, "reward": 0.4399881958961487, "reward_std": 0.5058347582817078, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19247499108314514, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5175000429153442, "step": 1631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 459.1875, "completions/mean_terminated_length": 459.1875, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.3264, "grad_norm": 0.11219805479049683, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18928941.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 652.3125, "completions/mean_terminated_length": 599.2142944335938, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.3266, "grad_norm": 0.13696733117103577, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18942898.0, "reward": 0.9550008177757263, "reward_std": 0.5874747633934021, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5350000262260437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 1633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 415.1875, "completions/mean_terminated_length": 415.1875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.3268, "grad_norm": 0.061986420303583145, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18953557.0, "reward": 0.8382818102836609, "reward_std": 0.1704999804496765, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6765625476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4750000238418579, "step": 1634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 702.375, "completions/mean_terminated_length": 656.4285888671875, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.327, "grad_norm": 0.10702627152204514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18970395.0, "reward": 0.611484944820404, "reward_std": 0.348372220993042, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4104687571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.550000011920929, "step": 1635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 342.1875, "completions/mean_terminated_length": 342.1875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.3272, "grad_norm": 0.0001412976998835802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18980478.0, "reward": 1.4989008903503418, "reward_std": 0.0004099227662663907, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977999925613403, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 346.375, "completions/mean_terminated_length": 346.375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.3274, "grad_norm": 0.1858743131160736, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18989796.0, "reward": 1.2416415214538574, "reward_std": 0.40442395210266113, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7957812547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 1637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 540.5625, "completions/mean_terminated_length": 540.5625, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.3276, "grad_norm": 0.05064713954925537, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19001885.0, "reward": 0.6026541590690613, "reward_std": 0.23942889273166656, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428062468767166, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 731.8125, "completions/mean_terminated_length": 634.4166870117188, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3278, "grad_norm": 0.17713306844234467, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19018362.0, "reward": 0.5813286900520325, "reward_std": 0.387748122215271, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4126562476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.40937501192092896, "step": 1639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 357.4375, "completions/mean_terminated_length": 357.4375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.328, "grad_norm": 0.07819916307926178, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19028209.0, "reward": 0.7442194223403931, "reward_std": 0.14509162306785583, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4884375035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 1640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 395.1875, "completions/mean_terminated_length": 395.1875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.3282, "grad_norm": 0.15500690042972565, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19038388.0, "reward": 1.3698418140411377, "reward_std": 0.3381863534450531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8646812438964844, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 1641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 640.0625, "completions/mean_terminated_length": 640.0625, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.3284, "grad_norm": 0.14987128973007202, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19052245.0, "reward": 0.9003416299819946, "reward_std": 0.4793291389942169, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4256812632083893, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 1642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 784.5625, "completions/mean_terminated_length": 675.727294921875, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.3286, "grad_norm": 0.11530791968107224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19068542.0, "reward": 0.880234956741333, "reward_std": 0.6910883188247681, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5729687213897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.590624988079071, "step": 1643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 322.4375, "completions/mean_terminated_length": 322.4375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.3288, "grad_norm": 0.1461624950170517, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19077045.0, "reward": 0.7675511837005615, "reward_std": 0.43676477670669556, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2851000130176544, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 395.0, "completions/mean_terminated_length": 395.0, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.329, "grad_norm": 0.08947014808654785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19088037.0, "reward": 0.5126447677612305, "reward_std": 0.1368953436613083, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08778749406337738, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 372.0, "completions/max_terminated_length": 372.0, "completions/mean_length": 299.0, "completions/mean_terminated_length": 299.0, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.3292, "grad_norm": 0.2296595424413681, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19096453.0, "reward": 1.0150322914123535, "reward_std": 0.5005450248718262, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5300624966621399, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 1646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 730.3125, "completions/mean_terminated_length": 688.357177734375, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.3294, "grad_norm": 0.08884897828102112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19112698.0, "reward": 0.658984899520874, "reward_std": 0.3684193193912506, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5054687261581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4281250238418579, "step": 1647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 534.4375, "completions/mean_terminated_length": 534.4375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.3296, "grad_norm": 0.039613064378499985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19125233.0, "reward": 0.6139726638793945, "reward_std": 0.15293848514556885, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22794374823570251, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8393749594688416, "step": 1648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 307.9375, "completions/mean_terminated_length": 307.9375, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.3298, "grad_norm": 0.129794180393219, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19134272.0, "reward": 0.7141917943954468, "reward_std": 0.3899278938770294, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24088124930858612, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 1649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 310.0, "completions/max_terminated_length": 310.0, "completions/mean_length": 208.875, "completions/mean_terminated_length": 208.875, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.33, "grad_norm": 0.4161655306816101, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19141118.0, "reward": 1.1927075386047363, "reward_std": 0.47012072801589966, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6979125142097473, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 1650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 593.375, "completions/mean_terminated_length": 593.375, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.3302, "grad_norm": 0.07559353858232498, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19154484.0, "reward": 0.8914071321487427, "reward_std": 0.30499669909477234, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6578125357627869, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5625, "step": 1651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 282.25, "completions/mean_terminated_length": 282.25, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.3304, "grad_norm": 0.13571399450302124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19162440.0, "reward": 1.323801040649414, "reward_std": 0.37701013684272766, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.835099995136261, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 376.125, "completions/mean_terminated_length": 376.125, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.3306, "grad_norm": 0.14599372446537018, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19172314.0, "reward": 0.8456259369850159, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3787499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 229.125, "completions/mean_terminated_length": 229.125, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.3308, "grad_norm": 0.2778639495372772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19179644.0, "reward": 1.0413135290145874, "reward_std": 0.48607099056243896, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5826249718666077, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 1654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 678.8125, "completions/mean_terminated_length": 629.5, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.331, "grad_norm": 0.09710310399532318, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19194217.0, "reward": 0.6246882677078247, "reward_std": 0.397464394569397, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24937500059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.793749988079071, "step": 1655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 395.125, "completions/mean_terminated_length": 395.125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.3312, "grad_norm": 0.09518016129732132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19204651.0, "reward": 0.6500540971755981, "reward_std": 0.26487991213798523, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2376062273979187, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8693749904632568, "step": 1656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 413.1875, "completions/mean_terminated_length": 413.1875, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.3314, "grad_norm": 0.08948244899511337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19215342.0, "reward": 1.4364852905273438, "reward_std": 0.23698069155216217, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9354687929153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 1657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 856.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 455.5625, "completions/mean_terminated_length": 455.5625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.3316, "grad_norm": 0.17593389749526978, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19227223.0, "reward": 0.7764853239059448, "reward_std": 0.3771911561489105, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.36546874046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656250238418579, "step": 1658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 345.0, "completions/mean_terminated_length": 345.0, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.3318, "grad_norm": 0.1173057034611702, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19236503.0, "reward": 0.7710133790969849, "reward_std": 0.32005950808525085, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.41702497005462646, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8025000095367432, "step": 1659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 431.5, "completions/mean_terminated_length": 431.5, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.332, "grad_norm": 0.14555951952934265, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19246943.0, "reward": 0.8438165187835693, "reward_std": 0.456093430519104, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3751312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 374.8125, "completions/mean_terminated_length": 374.8125, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.3322, "grad_norm": 0.06826528161764145, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19256492.0, "reward": 0.6382039785385132, "reward_std": 0.13205169141292572, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2764062285423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "step": 1661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 374.5, "completions/mean_terminated_length": 374.5, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.3324, "grad_norm": 0.16019803285598755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19266580.0, "reward": 0.7802071571350098, "reward_std": 0.4286435842514038, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3104124963283539, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 332.8125, "completions/mean_terminated_length": 332.8125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.3326, "grad_norm": 0.16290222108364105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19275793.0, "reward": 1.3207823038101196, "reward_std": 0.3830350637435913, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290625214576721, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 404.8125, "completions/mean_terminated_length": 404.8125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.3328, "grad_norm": 0.11627431958913803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19286350.0, "reward": 1.1367852687835693, "reward_std": 0.40733662247657776, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7735687494277954, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7206249833106995, "step": 1664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 680.75, "completions/mean_terminated_length": 680.75, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.333, "grad_norm": 0.12906219065189362, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19301066.0, "reward": 0.6432039737701416, "reward_std": 0.3134993314743042, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2864062488079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 1665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 430.625, "completions/mean_terminated_length": 391.0666809082031, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.3332, "grad_norm": 0.20580536127090454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19311380.0, "reward": 1.1080353260040283, "reward_std": 0.5369977355003357, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6535687446594238, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 1666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 530.0625, "completions/mean_terminated_length": 530.0625, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.3334, "grad_norm": 0.10578511655330658, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19326229.0, "reward": 1.0239073038101196, "reward_std": 0.4907395541667938, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5478125214576721, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 606.1875, "completions/mean_terminated_length": 578.3333740234375, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.3336, "grad_norm": 0.0750538781285286, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19339800.0, "reward": 0.6831260323524475, "reward_std": 0.38234415650367737, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3037499785423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.800000011920929, "step": 1668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 263.8125, "completions/mean_terminated_length": 263.8125, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.3338, "grad_norm": 0.17524053156375885, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19348277.0, "reward": 1.3164010047912598, "reward_std": 0.39381301403045654, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8202999830245972, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 1669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1004.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 624.8125, "completions/mean_terminated_length": 624.8125, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.334, "grad_norm": 0.10625714063644409, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19362962.0, "reward": 0.9519949555397034, "reward_std": 0.49904748797416687, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4664875268936157, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 1670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 475.125, "completions/mean_terminated_length": 438.5333557128906, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.3342, "grad_norm": 0.13652931153774261, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19374548.0, "reward": 0.7150007486343384, "reward_std": 0.37676864862442017, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3675000071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 1671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 337.0, "completions/mean_terminated_length": 337.0, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.3344, "grad_norm": 0.20182248950004578, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19383636.0, "reward": 1.0790977478027344, "reward_std": 0.4919913411140442, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5956937670707703, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 1672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 500.0, "completions/mean_terminated_length": 465.0666809082031, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.3346, "grad_norm": 0.031050754711031914, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19395620.0, "reward": 0.5455477237701416, "reward_std": 0.1837986260652542, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1535937488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.828125, "step": 1673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 690.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 527.375, "completions/mean_terminated_length": 527.375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.3348, "grad_norm": 0.05703090876340866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19408154.0, "reward": 1.439516544342041, "reward_std": 0.23753778636455536, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415313005447388, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 592.3125, "completions/mean_terminated_length": 563.5333862304688, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.335, "grad_norm": 0.03860105574131012, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19423887.0, "reward": 0.5126446485519409, "reward_std": 0.13689583539962769, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08778749406337738, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 1675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 640.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 483.875, "completions/mean_terminated_length": 483.875, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.3352, "grad_norm": 0.11478590220212936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19436029.0, "reward": 0.6778104305267334, "reward_std": 0.32527509331703186, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23061874508857727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9368749856948853, "step": 1676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 469.5625, "completions/mean_terminated_length": 432.60003662109375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.3354, "grad_norm": 0.16423435509204865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19447254.0, "reward": 0.7130447030067444, "reward_std": 0.42117810249328613, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.30108746886253357, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8712500333786011, "step": 1677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 268.8125, "completions/mean_terminated_length": 268.8125, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.3356, "grad_norm": 0.011656672693789005, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19455107.0, "reward": 0.5355695486068726, "reward_std": 0.020340682938694954, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07113693654537201, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.963562548160553, "step": 1678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 420.875, "completions/mean_terminated_length": 420.875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.3358, "grad_norm": 0.0671372041106224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19465425.0, "reward": 0.6495260000228882, "reward_std": 0.20829962193965912, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.299049973487854, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7425000071525574, "step": 1679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 273.0625, "completions/mean_terminated_length": 273.0625, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.336, "grad_norm": 0.00032356425072066486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19474258.0, "reward": 1.4993668794631958, "reward_std": 0.0006363937864080071, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998731255531311, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 1680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 822.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 660.5625, "completions/mean_terminated_length": 660.5625, "completions/min_length": 509.0, "completions/min_terminated_length": 509.0, "epoch": 0.3362, "grad_norm": 0.030774252489209175, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19488475.0, "reward": 0.597566545009613, "reward_std": 0.12124456465244293, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19513124227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8831250071525574, "step": 1681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 311.8125, "completions/mean_terminated_length": 311.8125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.3364, "grad_norm": 0.0002921012055594474, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19497496.0, "reward": 1.499035358428955, "reward_std": 0.0005109169869683683, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980687499046326, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 219.0, "completions/max_terminated_length": 219.0, "completions/mean_length": 168.5625, "completions/mean_terminated_length": 168.5625, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.3366, "grad_norm": 0.00026615781825967133, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19503553.0, "reward": 1.4998199939727783, "reward_std": 0.00041805318323895335, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9996374845504761, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9887499809265137, "step": 1683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 296.5625, "completions/mean_terminated_length": 296.5625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.3368, "grad_norm": 0.12664280831813812, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19512506.0, "reward": 1.3803073167800903, "reward_std": 0.3246067762374878, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8856124877929688, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 1684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 767.3125, "completions/mean_terminated_length": 750.2000122070312, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.337, "grad_norm": 0.0181423407047987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19528735.0, "reward": 0.524297833442688, "reward_std": 0.14977221190929413, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11109374463558197, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 1685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 297.125, "completions/mean_terminated_length": 297.125, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.3372, "grad_norm": 0.16439570486545563, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19537361.0, "reward": 1.2655572891235352, "reward_std": 0.4189576804637909, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7811124920845032, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 354.0, "completions/max_terminated_length": 354.0, "completions/mean_length": 305.3125, "completions/mean_terminated_length": 305.3125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.3374, "grad_norm": 0.17772342264652252, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19546262.0, "reward": 1.137941598892212, "reward_std": 0.48183828592300415, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.650881290435791, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 276.75, "completions/mean_terminated_length": 276.75, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.3376, "grad_norm": 0.09582504630088806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19554754.0, "reward": 0.605079174041748, "reward_std": 0.2386225014925003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 742.0625, "completions/mean_terminated_length": 613.9091186523438, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.3378, "grad_norm": 0.07489007711410522, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19571219.0, "reward": 0.3923163115978241, "reward_std": 0.28042659163475037, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09713125228881836, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 1689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 998.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 721.4375, "completions/mean_terminated_length": 721.4375, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.338, "grad_norm": 0.05870954319834709, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19586842.0, "reward": 0.7410944104194641, "reward_std": 0.19096475839614868, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4821874797344208, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6499999761581421, "step": 1690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 258.875, "completions/mean_terminated_length": 258.875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.3382, "grad_norm": 0.23243758082389832, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19594616.0, "reward": 1.1308484077453613, "reward_std": 0.4918922483921051, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6366937160491943, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 1691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 578.875, "completions/mean_terminated_length": 549.2000122070312, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.3384, "grad_norm": 0.06245030090212822, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19607782.0, "reward": 0.8097660541534424, "reward_std": 0.2602887749671936, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6820312738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.40937501192092896, "step": 1692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 374.5625, "completions/mean_terminated_length": 374.5625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.3386, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19618607.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 286.125, "completions/mean_terminated_length": 286.125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.3388, "grad_norm": 0.08334875851869583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19627201.0, "reward": 1.4397447109222412, "reward_std": 0.23759900033473969, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419875144958496, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 1694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 912.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 607.8125, "completions/mean_terminated_length": 607.8125, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.339, "grad_norm": 0.07732317596673965, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19640638.0, "reward": 0.7225509881973267, "reward_std": 0.3858368694782257, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25760000944137573, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 699.625, "completions/mean_terminated_length": 678.0000610351562, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.3392, "grad_norm": 0.09279657155275345, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19656104.0, "reward": 1.0606135129928589, "reward_std": 0.5031008720397949, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6837249994277954, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7737500071525574, "step": 1696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 667.875, "completions/mean_terminated_length": 644.1333618164062, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.3394, "grad_norm": 0.05734153464436531, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19670870.0, "reward": 0.9543753266334534, "reward_std": 0.315715491771698, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.783750057220459, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.2750000059604645, "step": 1697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 405.5, "completions/mean_terminated_length": 405.5, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.3396, "grad_norm": 0.18341360986232758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19681038.0, "reward": 1.0166666507720947, "reward_std": 0.49850353598594666, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5333312749862671, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 542.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 347.3125, "completions/mean_terminated_length": 347.3125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.3398, "grad_norm": 0.15030652284622192, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19691107.0, "reward": 0.6172605752944946, "reward_std": 0.35611453652381897, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17201873660087585, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9300000071525574, "step": 1699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 976.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 642.9375, "completions/mean_terminated_length": 642.9375, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.34, "grad_norm": 0.09254980087280273, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19705442.0, "reward": 0.8353761434555054, "reward_std": 0.4624219238758087, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3582499921321869, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 397.8125, "completions/mean_terminated_length": 397.8125, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.3402, "grad_norm": 0.14257773756980896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19715855.0, "reward": 1.230210542678833, "reward_std": 0.4253779947757721, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7729187607765198, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 1701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 480.3125, "completions/mean_terminated_length": 480.3125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.3404, "grad_norm": 0.11826235800981522, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19727204.0, "reward": 0.8303918838500977, "reward_std": 0.46584945917129517, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34828126430511475, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 1702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 671.375, "completions/mean_terminated_length": 590.0, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.3406, "grad_norm": 0.1608673632144928, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19742058.0, "reward": 0.8822661638259888, "reward_std": 0.5454391837120056, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6395312547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49687498807907104, "step": 1703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 209.0, "completions/max_terminated_length": 209.0, "completions/mean_length": 149.9375, "completions/mean_terminated_length": 149.9375, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "epoch": 0.3408, "grad_norm": 0.2871665954589844, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19747977.0, "reward": 1.4375293254852295, "reward_std": 0.24974071979522705, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9375561475753784, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9930000305175781, "step": 1704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 304.125, "completions/mean_terminated_length": 304.125, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.341, "grad_norm": 0.00010652383207343519, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19756635.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 577.4375, "completions/mean_terminated_length": 547.6666870117188, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.3412, "grad_norm": 0.09339207410812378, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19769874.0, "reward": 0.8060945272445679, "reward_std": 0.46382853388786316, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4246875047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.793749988079071, "step": 1706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 542.875, "completions/mean_terminated_length": 510.8000183105469, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.3414, "grad_norm": 0.14451412856578827, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19782352.0, "reward": 0.9434571862220764, "reward_std": 0.5373767614364624, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5119125247001648, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8700000047683716, "step": 1707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 470.9375, "completions/mean_terminated_length": 470.9375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.3416, "grad_norm": 0.022392362356185913, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19793695.0, "reward": 0.5630478858947754, "reward_std": 0.08482184261083603, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12609374523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 1708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 691.625, "completions/mean_terminated_length": 691.625, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.3418, "grad_norm": 0.10098889470100403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19808617.0, "reward": 1.2507822513580322, "reward_std": 0.3959769904613495, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8140624761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8887500166893005, "step": 1709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 565.0625, "completions/mean_terminated_length": 565.0625, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.342, "grad_norm": 0.048120152205228806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19821290.0, "reward": 0.7056257128715515, "reward_std": 0.12210121750831604, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4112499952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.75, "step": 1710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 457.125, "completions/mean_terminated_length": 419.3333435058594, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.3422, "grad_norm": 0.05956409126520157, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19833308.0, "reward": 0.5719482898712158, "reward_std": 0.3076070249080658, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1438937485218048, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 1711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 535.125, "completions/mean_terminated_length": 502.5333557128906, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.3424, "grad_norm": 0.12076558917760849, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19845742.0, "reward": 0.7738257646560669, "reward_std": 0.429238498210907, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.42264997959136963, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7462499737739563, "step": 1712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 389.4375, "completions/mean_terminated_length": 389.4375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.3426, "grad_norm": 0.0806545838713646, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19855653.0, "reward": 0.6447635889053345, "reward_std": 0.19597777724266052, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.28952500224113464, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7587499618530273, "step": 1713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 844.0625, "completions/mean_terminated_length": 762.2727661132812, "completions/min_length": 550.0, "completions/min_terminated_length": 550.0, "epoch": 0.3428, "grad_norm": 0.07348565757274628, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19873014.0, "reward": 0.5510162115097046, "reward_std": 0.4835837781429291, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28953126072883606, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5843750238418579, "step": 1714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 491.0, "completions/max_terminated_length": 491.0, "completions/mean_length": 371.9375, "completions/mean_terminated_length": 371.9375, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.343, "grad_norm": 0.0035866503603756428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19882725.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 473.5625, "completions/mean_terminated_length": 473.5625, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.3432, "grad_norm": 9.907145431498066e-05, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19894350.0, "reward": 1.4988821744918823, "reward_std": 0.0003586417296901345, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 684.875, "completions/mean_terminated_length": 636.4285888671875, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.3434, "grad_norm": 0.12801022827625275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19909548.0, "reward": 0.8279321193695068, "reward_std": 0.5283985137939453, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4683625102043152, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7462500333786011, "step": 1717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 986.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 625.8125, "completions/mean_terminated_length": 625.8125, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.3436, "grad_norm": 0.07624009996652603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19923657.0, "reward": 0.8608601093292236, "reward_std": 0.30929428339004517, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5967187881469727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6281249523162842, "step": 1718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 476.375, "completions/mean_terminated_length": 476.375, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.3438, "grad_norm": 0.1129673421382904, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19935407.0, "reward": 1.200338363647461, "reward_std": 0.45778799057006836, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7131750583648682, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 481.0, "completions/mean_terminated_length": 481.0, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.344, "grad_norm": 0.11060234159231186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19947023.0, "reward": 0.7322665452957153, "reward_std": 0.36306872963905334, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2770312428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 1720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 290.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 222.4375, "completions/mean_terminated_length": 222.4375, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.3442, "grad_norm": 0.3523350656032562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19954214.0, "reward": 1.0265604257583618, "reward_std": 0.48826950788497925, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5531187057495117, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 1721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 338.5625, "completions/mean_terminated_length": 338.5625, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.3444, "grad_norm": 0.00919286534190178, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19963743.0, "reward": 0.5390447378158569, "reward_std": 0.01773521862924099, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07808749377727509, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 235.5, "completions/mean_terminated_length": 235.5, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.3446, "grad_norm": 0.03869974985718727, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19971751.0, "reward": 0.5656167268753052, "reward_std": 0.10200396180152893, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13123124837875366, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9231250286102295, "step": 1723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 882.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 659.9375, "completions/mean_terminated_length": 659.9375, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.3448, "grad_norm": 0.13417790830135345, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19986694.0, "reward": 1.0237510204315186, "reward_std": 0.4905793368816376, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 436.125, "completions/mean_terminated_length": 436.125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.345, "grad_norm": 0.12234622240066528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19998200.0, "reward": 0.8845977783203125, "reward_std": 0.4008450210094452, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5191937685012817, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7956249713897705, "step": 1725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 538.875, "completions/mean_terminated_length": 538.875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.3452, "grad_norm": 0.10352206975221634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20011238.0, "reward": 0.7021417617797852, "reward_std": 0.342156320810318, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27928122878074646, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8818750381469727, "step": 1726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 625.5, "completions/mean_terminated_length": 568.5714721679688, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.3454, "grad_norm": 0.056735992431640625, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20024942.0, "reward": 0.6010165810585022, "reward_std": 0.3987921178340912, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20203125476837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.828125, "step": 1727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 317.0, "completions/max_terminated_length": 317.0, "completions/mean_length": 248.75, "completions/mean_terminated_length": 248.75, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.3456, "grad_norm": 0.13320013880729675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20032858.0, "reward": 1.4287385940551758, "reward_std": 0.24958477914333344, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9199749827384949, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "step": 1728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 434.625, "completions/mean_terminated_length": 395.3333435058594, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.3458, "grad_norm": 0.10746495425701141, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20043876.0, "reward": 0.6181728839874268, "reward_std": 0.36907076835632324, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17384375631809235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 1729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 408.5, "completions/mean_terminated_length": 408.5, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.346, "grad_norm": 0.1948765367269516, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20054684.0, "reward": 0.9387233257293701, "reward_std": 0.5111986994743347, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.43994373083114624, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9943749904632568, "step": 1730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 586.625, "completions/mean_terminated_length": 586.625, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.3462, "grad_norm": 0.007285818923264742, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20067830.0, "reward": 0.5315513610839844, "reward_std": 0.028870603069663048, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06309999525547028, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 1731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 510.5625, "completions/mean_terminated_length": 476.3333435058594, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.3464, "grad_norm": 0.1766403466463089, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20080127.0, "reward": 0.9330352544784546, "reward_std": 0.5329147577285767, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4910687804222107, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 1732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 439.25, "completions/mean_terminated_length": 439.25, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.3466, "grad_norm": 0.09727317839860916, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20092051.0, "reward": 0.6350008249282837, "reward_std": 0.2359360158443451, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.33249998092651367, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7312500476837158, "step": 1733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 430.3125, "completions/mean_terminated_length": 430.3125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.3468, "grad_norm": 0.17428886890411377, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20102728.0, "reward": 1.1090384721755981, "reward_std": 0.5364503264427185, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6555749773979187, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8912500143051147, "step": 1734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 362.3125, "completions/mean_terminated_length": 362.3125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.347, "grad_norm": 0.00021156604634597898, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20112301.0, "reward": 1.4989697933197021, "reward_std": 0.0004720082797575742, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979375004768372, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 1735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 418.9375, "completions/mean_terminated_length": 378.6000061035156, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.3472, "grad_norm": 0.1545829176902771, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20123324.0, "reward": 1.2193446159362793, "reward_std": 0.4795311689376831, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7511874437332153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8212500214576721, "step": 1736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 586.0625, "completions/mean_terminated_length": 586.0625, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.3474, "grad_norm": 0.07456132024526596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20136605.0, "reward": 0.8801571130752563, "reward_std": 0.3463267385959625, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5728124380111694, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.71875, "step": 1737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 493.5, "completions/mean_terminated_length": 493.5, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.3476, "grad_norm": 0.13014601171016693, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20148517.0, "reward": 0.9613291025161743, "reward_std": 0.48964908719062805, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48515623807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 402.5625, "completions/mean_terminated_length": 402.5625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.3478, "grad_norm": 0.12141704559326172, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20158814.0, "reward": 0.9019541144371033, "reward_std": 0.4775867462158203, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42890626192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 748.875, "completions/mean_terminated_length": 730.5333862304688, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.348, "grad_norm": 0.05444381758570671, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20174876.0, "reward": 0.6355445981025696, "reward_std": 0.30574873089790344, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2710874676704407, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7962499856948853, "step": 1740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 468.8125, "completions/mean_terminated_length": 468.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3482, "grad_norm": 0.16024379432201385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20186777.0, "reward": 1.0238134860992432, "reward_std": 0.4912717640399933, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.547624945640564, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 407.125, "completions/mean_terminated_length": 407.125, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.3484, "grad_norm": 0.14901471138000488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20197355.0, "reward": 1.0831258296966553, "reward_std": 0.48673015832901, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6037499904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 436.0, "completions/max_terminated_length": 436.0, "completions/mean_length": 266.125, "completions/mean_terminated_length": 266.125, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.3486, "grad_norm": 0.12998539209365845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20205149.0, "reward": 1.437629222869873, "reward_std": 0.2473815530538559, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9377562999725342, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 1743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 420.8125, "completions/mean_terminated_length": 420.8125, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.3488, "grad_norm": 0.10139638930559158, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20216186.0, "reward": 1.3802978992462158, "reward_std": 0.3246031403541565, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855937719345093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 408.5625, "completions/mean_terminated_length": 408.5625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.349, "grad_norm": 0.1418968290090561, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20226435.0, "reward": 1.0808415412902832, "reward_std": 0.4897860586643219, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5991812944412231, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 1745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 585.5625, "completions/mean_terminated_length": 585.5625, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.3492, "grad_norm": 0.05891801416873932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20239660.0, "reward": 0.817031741142273, "reward_std": 0.25643518567085266, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6965625286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.41875001788139343, "step": 1746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 305.6875, "completions/mean_terminated_length": 305.6875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.3494, "grad_norm": 0.28697922825813293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20248935.0, "reward": 1.4058449268341064, "reward_std": 0.3748917579650879, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9366874694824219, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 1747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 592.5, "completions/mean_terminated_length": 530.857177734375, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.3496, "grad_norm": 0.06854240596294403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20262463.0, "reward": 0.8367974162101746, "reward_std": 0.3797161877155304, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.7360937595367432, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.296875, "step": 1748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 267.1875, "completions/mean_terminated_length": 267.1875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.3498, "grad_norm": 0.009269734844565392, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20271138.0, "reward": 0.537857174873352, "reward_std": 0.016918901354074478, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07571249455213547, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 1749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 301.3125, "completions/mean_terminated_length": 301.3125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.35, "grad_norm": 0.18582071363925934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20280391.0, "reward": 1.2021042108535767, "reward_std": 0.45493701100349426, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7167062759399414, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 348.8125, "completions/mean_terminated_length": 348.8125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.3502, "grad_norm": 0.1407102793455124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20290196.0, "reward": 1.1660072803497314, "reward_std": 0.4535534679889679, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7070125341415405, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8887499570846558, "step": 1751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 557.1875, "completions/mean_terminated_length": 490.5000305175781, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.3504, "grad_norm": 0.1193937286734581, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20303159.0, "reward": 0.6815632581710815, "reward_std": 0.4572458863258362, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3006249964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7999999523162842, "step": 1752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 298.8125, "completions/mean_terminated_length": 298.8125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.3506, "grad_norm": 0.10283797979354858, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20311668.0, "reward": 0.8650788068771362, "reward_std": 0.18881474435329437, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7301562428474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.37187498807907104, "step": 1753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 707.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 495.75, "completions/mean_terminated_length": 495.75, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.3508, "grad_norm": 0.03756462410092354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20324176.0, "reward": 0.5778728723526001, "reward_std": 0.12252414971590042, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15574374794960022, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8881250023841858, "step": 1754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 481.375, "completions/mean_terminated_length": 481.375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.351, "grad_norm": 0.07824522256851196, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20335670.0, "reward": 0.7708602547645569, "reward_std": 0.3311923146247864, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4167187511920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 1755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 368.6875, "completions/mean_terminated_length": 368.6875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.3512, "grad_norm": 0.12293373048305511, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20345521.0, "reward": 1.3805854320526123, "reward_std": 0.32471543550491333, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8861687779426575, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 228.1875, "completions/mean_terminated_length": 228.1875, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.3514, "grad_norm": 0.0029144950676709414, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20352772.0, "reward": 1.497641921043396, "reward_std": 0.007725962437689304, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9952812194824219, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 1757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 379.5625, "completions/mean_terminated_length": 379.5625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.3516, "grad_norm": 0.11718539893627167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20362413.0, "reward": 1.193713665008545, "reward_std": 0.46825677156448364, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6999250054359436, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 786.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 419.4375, "completions/mean_terminated_length": 419.4375, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.3518, "grad_norm": 0.09777798503637314, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20373220.0, "reward": 0.700391411781311, "reward_std": 0.18930251896381378, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4007812440395355, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6968749761581421, "step": 1759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 299.0, "completions/max_terminated_length": 299.0, "completions/mean_length": 241.8125, "completions/mean_terminated_length": 241.8125, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.352, "grad_norm": 0.125198632478714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20380657.0, "reward": 1.4376355409622192, "reward_std": 0.2473832666873932, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9377687573432922, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750667572021, "step": 1760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 516.5625, "completions/mean_terminated_length": 516.5625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.3522, "grad_norm": 0.0943581685423851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20393306.0, "reward": 0.7935945987701416, "reward_std": 0.2984123229980469, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5246875286102295, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5874999761581421, "step": 1761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 577.9375, "completions/mean_terminated_length": 514.2142944335938, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.3524, "grad_norm": 0.12749923765659332, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20406105.0, "reward": 0.6228914856910706, "reward_std": 0.3649408519268036, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3082812428474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6781250238418579, "step": 1762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 541.3125, "completions/mean_terminated_length": 541.3125, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.3526, "grad_norm": 0.096162348985672, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20418814.0, "reward": 1.2611477375030518, "reward_std": 0.42479267716407776, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.772293746471405, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 1763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 729.625, "completions/mean_terminated_length": 710.0000610351562, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.3528, "grad_norm": 0.09693168103694916, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20434856.0, "reward": 0.9346885681152344, "reward_std": 0.5336349010467529, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49437499046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 1764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 588.625, "completions/mean_terminated_length": 588.625, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.353, "grad_norm": 0.005697743035852909, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20447890.0, "reward": 0.5281386375427246, "reward_std": 0.0200456902384758, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05627499520778656, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 1765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 329.875, "completions/mean_terminated_length": 329.875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.3532, "grad_norm": 0.5192830562591553, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20456976.0, "reward": 1.2841103076934814, "reward_std": 0.4753490686416626, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8182188272476196, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 832.0, "completions/mean_terminated_length": 804.5714721679688, "completions/min_length": 593.0, "completions/min_terminated_length": 593.0, "epoch": 0.3534, "grad_norm": 0.03897643834352493, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20474848.0, "reward": 0.5171039700508118, "reward_std": 0.2400462031364441, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15920624136924744, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7643749713897705, "step": 1767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 378.9375, "completions/mean_terminated_length": 378.9375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.3536, "grad_norm": 0.15035538375377655, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20485295.0, "reward": 1.3856258392333984, "reward_std": 0.30955278873443604, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8962500095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 1768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 437.8125, "completions/mean_terminated_length": 437.8125, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.3538, "grad_norm": 0.10405196249485016, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20496540.0, "reward": 1.2162511348724365, "reward_std": 0.4014298617839813, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8075000047683716, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 1769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 271.0, "completions/max_terminated_length": 271.0, "completions/mean_length": 219.1875, "completions/mean_terminated_length": 219.1875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.354, "grad_norm": 0.270729124546051, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20504047.0, "reward": 1.0677043199539185, "reward_std": 0.5056737065315247, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5729062557220459, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 1770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 631.375, "completions/mean_terminated_length": 605.2000122070312, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.3542, "grad_norm": 0.020217986777424812, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20517845.0, "reward": 0.4986480474472046, "reward_std": 0.13469591736793518, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05979374796152115, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 1771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 414.5625, "completions/mean_terminated_length": 414.5625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.3544, "grad_norm": 0.07548654824495316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20528350.0, "reward": 0.6887384653091431, "reward_std": 0.3301146626472473, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25247499346733093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 1772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 663.875, "completions/mean_terminated_length": 612.4285888671875, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.3546, "grad_norm": 0.06748198717832565, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20542764.0, "reward": 0.7197661399841309, "reward_std": 0.36150023341178894, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5020312666893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.565625011920929, "step": 1773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 550.3125, "completions/mean_terminated_length": 518.7333374023438, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.3548, "grad_norm": 0.1075824573636055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20555281.0, "reward": 1.221351146697998, "reward_std": 0.5107218027114868, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7552000284194946, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 1774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 485.8125, "completions/mean_terminated_length": 485.8125, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.355, "grad_norm": 0.006378720980137587, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20567630.0, "reward": 0.5461759567260742, "reward_std": 0.019841261208057404, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09234999120235443, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524937868118286, "step": 1775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 371.0, "completions/max_terminated_length": 371.0, "completions/mean_length": 297.9375, "completions/mean_terminated_length": 297.9375, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.3552, "grad_norm": 0.20737138390541077, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20576221.0, "reward": 1.0801572799682617, "reward_std": 0.49050796031951904, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5978125333786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 1776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 569.125, "completions/mean_terminated_length": 569.125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3554, "grad_norm": 0.08685032278299332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20589423.0, "reward": 0.6939854025840759, "reward_std": 0.3333006799221039, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2629687488079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 1777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 552.25, "completions/mean_terminated_length": 520.800048828125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.3556, "grad_norm": 0.18060900270938873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20602147.0, "reward": 1.12404465675354, "reward_std": 0.5144134759902954, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.685587465763092, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 1778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 531.375, "completions/mean_terminated_length": 531.375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.3558, "grad_norm": 0.11898300796747208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20614521.0, "reward": 1.2834761142730713, "reward_std": 0.40081411600112915, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8169499635696411, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 1779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 373.0625, "completions/mean_terminated_length": 373.0625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.356, "grad_norm": 0.005526768509298563, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20624010.0, "reward": 0.5439009666442871, "reward_std": 0.013252700679004192, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08779999613761902, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 586.0625, "completions/mean_terminated_length": 556.86669921875, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.3562, "grad_norm": 0.03637487813830376, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20638459.0, "reward": 0.5492008328437805, "reward_std": 0.18888278305530548, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16089999675750732, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8287500143051147, "step": 1781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 954.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 476.0, "completions/mean_terminated_length": 476.0, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.3564, "grad_norm": 0.08304325491189957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20649803.0, "reward": 1.3207790851593018, "reward_std": 0.38303348422050476, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290562629699707, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 361.75, "completions/mean_terminated_length": 361.75, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.3566, "grad_norm": 0.17785769701004028, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20659447.0, "reward": 1.2590605020523071, "reward_std": 0.429418683052063, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7681187391281128, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 1783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 344.0, "completions/mean_terminated_length": 344.0, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.3568, "grad_norm": 0.10005691647529602, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20668967.0, "reward": 1.3829697370529175, "reward_std": 0.316731333732605, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8909375071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 424.0625, "completions/mean_terminated_length": 424.0625, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.357, "grad_norm": 0.13725689053535461, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20679624.0, "reward": 1.0145196914672852, "reward_std": 0.569068968296051, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5915374755859375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7237499952316284, "step": 1785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 575.75, "completions/mean_terminated_length": 545.86669921875, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.3572, "grad_norm": 0.06836657226085663, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20692804.0, "reward": 0.8703131675720215, "reward_std": 0.2782173156738281, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8031250238418579, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.19375000894069672, "step": 1786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 267.0, "completions/max_terminated_length": 267.0, "completions/mean_length": 199.0, "completions/mean_terminated_length": 199.0, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.3574, "grad_norm": 0.2714358866214752, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20699780.0, "reward": 1.3770947456359863, "reward_std": 0.3329862952232361, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8791874647140503, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 248.0, "completions/max_terminated_length": 248.0, "completions/mean_length": 198.5, "completions/mean_terminated_length": 198.5, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.3576, "grad_norm": 0.21361860632896423, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20706844.0, "reward": 1.378110647201538, "reward_std": 0.331398069858551, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8812187910079956, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 1788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 720.5, "completions/mean_terminated_length": 677.1428833007812, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.3578, "grad_norm": 0.054942552000284195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20722644.0, "reward": 0.7514848709106445, "reward_std": 0.3398520052433014, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6279687881469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3531250059604645, "step": 1789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 420.5625, "completions/mean_terminated_length": 420.5625, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.358, "grad_norm": 0.08233615756034851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20732941.0, "reward": 1.320766568183899, "reward_std": 0.38302722573280334, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290312886238098, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 1790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 553.625, "completions/mean_terminated_length": 553.625, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3582, "grad_norm": 0.013663538731634617, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20746071.0, "reward": 0.5771102905273438, "reward_std": 0.06980278342962265, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15421874821186066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 1791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 476.1875, "completions/mean_terminated_length": 476.1875, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.3584, "grad_norm": 0.10152240842580795, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20757354.0, "reward": 0.7892197370529175, "reward_std": 0.4234236478805542, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.328437477350235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 638.0625, "completions/mean_terminated_length": 638.0625, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.3586, "grad_norm": 0.11697975546121597, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20771179.0, "reward": 1.1431009769439697, "reward_std": 0.47548019886016846, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6611999869346619, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 600.3125, "completions/mean_terminated_length": 600.3125, "completions/min_length": 457.0, "completions/min_terminated_length": 457.0, "epoch": 0.3588, "grad_norm": 0.08274485170841217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20784832.0, "reward": 1.2595667839050293, "reward_std": 0.429720401763916, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691312432289124, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 1794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 653.1875, "completions/mean_terminated_length": 628.4666748046875, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.359, "grad_norm": 0.039382029324769974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20799427.0, "reward": 0.589607298374176, "reward_std": 0.24380488693714142, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24171249568462372, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7337499856948853, "step": 1795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 453.9375, "completions/mean_terminated_length": 453.9375, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.3592, "grad_norm": 0.12326520681381226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20810786.0, "reward": 0.6251198053359985, "reward_std": 0.25579842925071716, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18773749470710754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 1796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 874.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 524.0, "completions/mean_terminated_length": 524.0, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.3594, "grad_norm": 0.05536390841007233, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20826258.0, "reward": 0.8956255912780762, "reward_std": 0.18696537613868713, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7912499904632568, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.2874999940395355, "step": 1797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 869.0, "completions/max_terminated_length": 869.0, "completions/mean_length": 614.8125, "completions/mean_terminated_length": 614.8125, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.3596, "grad_norm": 0.11808573454618454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20839807.0, "reward": 1.1674227714538574, "reward_std": 0.45495885610580444, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7098437547683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9031249284744263, "step": 1798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 563.125, "completions/mean_terminated_length": 563.125, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.3598, "grad_norm": 0.05888618528842926, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20852289.0, "reward": 0.6110915541648865, "reward_std": 0.2373097985982895, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15968123078346252, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 1799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 548.1875, "completions/mean_terminated_length": 548.1875, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.36, "grad_norm": 0.13190977275371552, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20865684.0, "reward": 0.9082819223403931, "reward_std": 0.38929787278175354, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5665625333786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 1800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 580.0, "completions/mean_terminated_length": 580.0, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.3602, "grad_norm": 0.00405677268281579, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20879012.0, "reward": 0.5485947728157043, "reward_std": 0.017349589616060257, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 624.6875, "completions/mean_terminated_length": 598.0667114257812, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.3604, "grad_norm": 0.07098111510276794, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20893023.0, "reward": 0.4935133755207062, "reward_std": 0.1999364197254181, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11202500015497208, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8743749856948853, "step": 1802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 371.875, "completions/mean_terminated_length": 371.875, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.3606, "grad_norm": 0.41931959986686707, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20902941.0, "reward": 0.9504572153091431, "reward_std": 0.5914919376373291, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5259125232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 1803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 819.375, "completions/mean_terminated_length": 805.7333984375, "completions/min_length": 664.0, "completions/min_terminated_length": 664.0, "epoch": 0.3608, "grad_norm": 0.0930158793926239, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20920563.0, "reward": 0.7089069485664368, "reward_std": 0.4012264907360077, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4178125262260437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6437499523162842, "step": 1804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 813.8125, "completions/mean_terminated_length": 799.800048828125, "completions/min_length": 580.0, "completions/min_terminated_length": 580.0, "epoch": 0.361, "grad_norm": 0.07267365604639053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20937648.0, "reward": 0.7646884322166443, "reward_std": 0.46275392174720764, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.34187501668930054, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.875, "step": 1805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 304.375, "completions/mean_terminated_length": 304.375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.3612, "grad_norm": 0.004769457969814539, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20946150.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 607.25, "completions/mean_terminated_length": 547.7142944335938, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.3614, "grad_norm": 0.14245149493217468, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20959786.0, "reward": 1.0694009065628052, "reward_std": 0.594636857509613, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6388000249862671, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8475000262260437, "step": 1807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 473.75, "completions/mean_terminated_length": 473.75, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.3616, "grad_norm": 0.14379307627677917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20971366.0, "reward": 1.2590415477752686, "reward_std": 0.42940735816955566, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7680813074111938, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 410.0, "completions/max_terminated_length": 410.0, "completions/mean_length": 318.9375, "completions/mean_terminated_length": 318.9375, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.3618, "grad_norm": 0.08157702535390854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20980517.0, "reward": 0.5952229499816895, "reward_std": 0.24233025312423706, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.127943754196167, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 935.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 642.0, "completions/mean_terminated_length": 642.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.362, "grad_norm": 0.007214306388050318, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20994245.0, "reward": 0.5461697578430176, "reward_std": 0.019856736063957214, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09233749657869339, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 427.125, "completions/mean_terminated_length": 427.125, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.3622, "grad_norm": 0.1730675846338272, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21004823.0, "reward": 1.2000665664672852, "reward_std": 0.457598477602005, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7126312851905823, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 852.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 551.8125, "completions/mean_terminated_length": 551.8125, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.3624, "grad_norm": 0.21243788301944733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21018516.0, "reward": 1.1047353744506836, "reward_std": 0.540769100189209, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6469687819480896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8993750214576721, "step": 1812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 501.0625, "completions/mean_terminated_length": 501.0625, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.3626, "grad_norm": 0.13903921842575073, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21030965.0, "reward": 1.0862321853637695, "reward_std": 0.4836679697036743, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.609962522983551, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 588.5, "completions/mean_terminated_length": 588.5, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.3628, "grad_norm": 0.09228802472352982, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21044173.0, "reward": 0.6933602094650269, "reward_std": 0.25774750113487244, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.32421875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8031249642372131, "step": 1814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 602.125, "completions/mean_terminated_length": 541.857177734375, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.363, "grad_norm": 0.11963871866464615, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21057519.0, "reward": 0.7718352675437927, "reward_std": 0.5373228192329407, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3561687469482422, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8456250429153442, "step": 1815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 288.6875, "completions/mean_terminated_length": 288.6875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.3632, "grad_norm": 0.2850446105003357, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21066506.0, "reward": 0.9585011601448059, "reward_std": 0.4927870035171509, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4794999957084656, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 374.4375, "completions/mean_terminated_length": 331.13336181640625, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.3634, "grad_norm": 0.15218371152877808, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21076465.0, "reward": 0.6433354020118713, "reward_std": 0.3785553276538849, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22416874766349792, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8618749976158142, "step": 1817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 382.0, "completions/mean_terminated_length": 382.0, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.3636, "grad_norm": 0.2465100735425949, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21086577.0, "reward": 1.284279227256775, "reward_std": 0.4754303991794586, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8185563087463379, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 1818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 949.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 665.25, "completions/mean_terminated_length": 665.25, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.3638, "grad_norm": 0.11129272729158401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21100821.0, "reward": 0.7023322582244873, "reward_std": 0.32378751039505005, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27966248989105225, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 1819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 366.75, "completions/mean_terminated_length": 366.75, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.364, "grad_norm": 0.00016158007201738656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21111041.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 466.6875, "completions/mean_terminated_length": 429.5333557128906, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.3642, "grad_norm": 0.14039063453674316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21122252.0, "reward": 0.8234226703643799, "reward_std": 0.45870909094810486, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4593437612056732, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7793750166893005, "step": 1821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 798.375, "completions/mean_terminated_length": 746.3077392578125, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.3644, "grad_norm": 0.10898207873106003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21139698.0, "reward": 0.8018757700920105, "reward_std": 0.5936394929885864, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41624999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 1822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 692.3125, "completions/mean_terminated_length": 670.2000122070312, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.3646, "grad_norm": 0.06705241650342941, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21155239.0, "reward": 0.8321098685264587, "reward_std": 0.28680458664894104, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7267187833786011, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 1823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 676.0625, "completions/mean_terminated_length": 626.357177734375, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.3648, "grad_norm": 0.04671965539455414, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21170488.0, "reward": 0.511469841003418, "reward_std": 0.24823644757270813, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14793750643730164, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7700000405311584, "step": 1824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 558.375, "completions/mean_terminated_length": 527.3333740234375, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.365, "grad_norm": 0.18080966174602509, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21183134.0, "reward": 1.0455479621887207, "reward_std": 0.5464581847190857, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5910937786102295, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 1825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 789.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 536.5, "completions/mean_terminated_length": 536.5, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.3652, "grad_norm": 0.15320467948913574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21195894.0, "reward": 1.2101447582244873, "reward_std": 0.4432949125766754, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7327874898910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 1826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 650.75, "completions/mean_terminated_length": 597.4285888671875, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.3654, "grad_norm": 0.07326626032590866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21210050.0, "reward": 0.531875729560852, "reward_std": 0.2391531616449356, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1887499988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 1827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 366.9375, "completions/mean_terminated_length": 366.9375, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.3656, "grad_norm": 0.22547774016857147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21219697.0, "reward": 0.833860456943512, "reward_std": 0.46409446001052856, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3552187383174896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 1828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 492.3125, "completions/mean_terminated_length": 456.86669921875, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.3658, "grad_norm": 0.07944168150424957, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21231318.0, "reward": 0.5546883344650269, "reward_std": 0.17939114570617676, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.171875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 1829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 259.0625, "completions/mean_terminated_length": 259.0625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.366, "grad_norm": 0.20772911608219147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21239287.0, "reward": 1.3469511270523071, "reward_std": 0.3392617404460907, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8813999891281128, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 1830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 331.4375, "completions/mean_terminated_length": 331.4375, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.3662, "grad_norm": 0.19343486428260803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21249150.0, "reward": 0.9607571363449097, "reward_std": 0.4901149272918701, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48401251435279846, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.3664, "grad_norm": 0.1851320117712021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21258752.0, "reward": 1.0118447542190552, "reward_std": 0.5035395622253418, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.523687481880188, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 1832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 799.375, "completions/mean_terminated_length": 697.2727661132812, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.3666, "grad_norm": 0.13499881327152252, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21275590.0, "reward": 0.5859379768371582, "reward_std": 0.46383872628211975, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.421875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.40625, "step": 1833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 427.5625, "completions/mean_terminated_length": 427.5625, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.3668, "grad_norm": 0.10860760509967804, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21286527.0, "reward": 1.3800791501998901, "reward_std": 0.3245174288749695, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851562738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 669.125, "completions/mean_terminated_length": 645.4666748046875, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.367, "grad_norm": 0.05221838876605034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21301329.0, "reward": 0.7166413068771362, "reward_std": 0.3135156035423279, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4332812428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.690625011920929, "step": 1835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 353.9375, "completions/mean_terminated_length": 353.9375, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.3672, "grad_norm": 0.17866238951683044, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21310800.0, "reward": 1.3237476348876953, "reward_std": 0.42904940247535706, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8349937200546265, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8868750333786011, "step": 1836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 947.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 540.25, "completions/mean_terminated_length": 540.25, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.3674, "grad_norm": 0.13893860578536987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21323332.0, "reward": 0.9550759196281433, "reward_std": 0.41058439016342163, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5976499915122986, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7899999618530273, "step": 1837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 641.125, "completions/mean_terminated_length": 615.6000366210938, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.3676, "grad_norm": 0.10088114440441132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21337078.0, "reward": 0.8703916072845459, "reward_std": 0.5208911895751953, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4282812476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 1838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 446.25, "completions/mean_terminated_length": 446.25, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.3678, "grad_norm": 0.17000505328178406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21348410.0, "reward": 0.9645072221755981, "reward_std": 0.4868839979171753, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4915125072002411, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 1839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 377.8125, "completions/mean_terminated_length": 377.8125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.368, "grad_norm": 0.1792379468679428, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21358503.0, "reward": 1.2020915746688843, "reward_std": 0.4549282193183899, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7166812419891357, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 1840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 337.75, "completions/mean_terminated_length": 337.75, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.3682, "grad_norm": 0.28649476170539856, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21367987.0, "reward": 0.9571728706359863, "reward_std": 0.4936707019805908, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47684377431869507, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 724.0, "completions/mean_terminated_length": 654.7692260742188, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.3684, "grad_norm": 0.04964436590671539, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21384707.0, "reward": 0.43919456005096436, "reward_std": 0.21838711202144623, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06588749587535858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7787500023841858, "step": 1842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 456.0625, "completions/mean_terminated_length": 456.0625, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.3686, "grad_norm": 0.12721367180347443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21398500.0, "reward": 0.9050009250640869, "reward_std": 0.47499996423721313, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4350000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 815.9375, "completions/mean_terminated_length": 786.2142944335938, "completions/min_length": 554.0, "completions/min_terminated_length": 554.0, "epoch": 0.3688, "grad_norm": 0.07842020690441132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21416547.0, "reward": 0.6030474901199341, "reward_std": 0.395749568939209, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3310937285423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 1844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 326.375, "completions/mean_terminated_length": 326.375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.369, "grad_norm": 0.11924377083778381, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21425385.0, "reward": 0.6905353665351868, "reward_std": 0.33080777525901794, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25606876611709595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9193750023841858, "step": 1845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 434.9375, "completions/mean_terminated_length": 434.9375, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.3692, "grad_norm": 0.16372869908809662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21435816.0, "reward": 1.0717949867248535, "reward_std": 0.5007372498512268, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5810874700546265, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 1846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 921.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 663.625, "completions/mean_terminated_length": 663.625, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.3694, "grad_norm": 0.04277573153376579, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21452082.0, "reward": 0.8558599352836609, "reward_std": 0.16072258353233337, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7117187976837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4281250238418579, "step": 1847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 562.125, "completions/mean_terminated_length": 562.125, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.3696, "grad_norm": 0.09531713277101517, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21465732.0, "reward": 0.7293758392333984, "reward_std": 0.3188769221305847, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.33375000953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.862500011920929, "step": 1848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 657.6875, "completions/mean_terminated_length": 657.6875, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.3698, "grad_norm": 0.020399795845150948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21480511.0, "reward": 0.5931260585784912, "reward_std": 0.10889759659767151, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1862500011920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8875000476837158, "step": 1849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 807.625, "completions/mean_terminated_length": 735.5, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.37, "grad_norm": 0.06474729627370834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21497625.0, "reward": 0.5221630930900574, "reward_std": 0.4068160951137543, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23182502388954163, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6162500381469727, "step": 1850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 799.75, "completions/mean_terminated_length": 697.8181762695312, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.3702, "grad_norm": 0.07362043857574463, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21515781.0, "reward": 0.38106951117515564, "reward_std": 0.2699670195579529, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0746375024318695, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6474999785423279, "step": 1851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 680.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 485.3125, "completions/mean_terminated_length": 485.3125, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.3704, "grad_norm": 0.13698571920394897, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21527802.0, "reward": 0.6757825016975403, "reward_std": 0.3331797122955322, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2265625, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 1852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 249.0, "completions/max_terminated_length": 249.0, "completions/mean_length": 202.0625, "completions/mean_terminated_length": 202.0625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.3706, "grad_norm": 0.0005075075896456838, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21535259.0, "reward": 1.4996013641357422, "reward_std": 0.0005925291916355491, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991999864578247, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 1853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 443.125, "completions/mean_terminated_length": 404.4000244140625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.3708, "grad_norm": 0.16609670221805573, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21546109.0, "reward": 0.814454197883606, "reward_std": 0.49620765447616577, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37890625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 1854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 840.0, "completions/max_terminated_length": 840.0, "completions/mean_length": 407.625, "completions/mean_terminated_length": 407.625, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.371, "grad_norm": 0.20243187248706818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21556503.0, "reward": 0.9001447558403015, "reward_std": 0.4790666103363037, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42528748512268066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 1855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 292.625, "completions/mean_terminated_length": 292.625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.3712, "grad_norm": 0.0034691228065639734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21564817.0, "reward": 0.5111323595046997, "reward_std": 0.010586373507976532, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02226249873638153, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9887500405311584, "step": 1856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 679.8125, "completions/mean_terminated_length": 630.6428833007812, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.3714, "grad_norm": 0.07136143743991852, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21579662.0, "reward": 0.5710914134979248, "reward_std": 0.33726149797439575, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20468124747276306, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7681249380111694, "step": 1857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 519.9375, "completions/mean_terminated_length": 519.9375, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.3716, "grad_norm": 0.12292156368494034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21591677.0, "reward": 1.041691780090332, "reward_std": 0.48647642135620117, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5833812952041626, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9206250309944153, "step": 1858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 280.625, "completions/mean_terminated_length": 280.625, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.3718, "grad_norm": 0.4223698675632477, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21599895.0, "reward": 0.903251051902771, "reward_std": 0.47722727060317993, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43150001764297485, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 641.125, "completions/mean_terminated_length": 615.6000366210938, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.372, "grad_norm": 0.13465559482574463, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21613881.0, "reward": 0.953113317489624, "reward_std": 0.5294548273086548, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5312249660491943, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 1860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 313.875, "completions/mean_terminated_length": 313.875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.3722, "grad_norm": 0.0002615471603348851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21623255.0, "reward": 1.499119758605957, "reward_std": 0.0005663654883392155, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982374906539917, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 1861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 635.625, "completions/mean_terminated_length": 546.0, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.3724, "grad_norm": 0.062921442091465, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21637041.0, "reward": 0.5034257769584656, "reward_std": 0.34388473629951477, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1318499892950058, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7737500071525574, "step": 1862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 538.8125, "completions/mean_terminated_length": 538.8125, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.3726, "grad_norm": 0.12201311439275742, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21649886.0, "reward": 1.0209228992462158, "reward_std": 0.49409109354019165, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5418437719345093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 375.5625, "completions/mean_terminated_length": 375.5625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.3728, "grad_norm": 0.21128329634666443, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21660487.0, "reward": 1.1433477401733398, "reward_std": 0.46969470381736755, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6616937518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 1864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 329.9375, "completions/mean_terminated_length": 329.9375, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.373, "grad_norm": 0.11951997876167297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21669990.0, "reward": 1.3803040981292725, "reward_std": 0.3246055841445923, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8856062889099121, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 1865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 866.1875, "completions/mean_terminated_length": 771.5, "completions/min_length": 554.0, "completions/min_terminated_length": 554.0, "epoch": 0.3732, "grad_norm": 0.1129714772105217, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21687961.0, "reward": 0.4865628480911255, "reward_std": 0.41918325424194336, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.34812498092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.35624998807907104, "step": 1866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 527.0625, "completions/mean_terminated_length": 493.933349609375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.3734, "grad_norm": 0.1531400829553604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21700410.0, "reward": 0.8020321726799011, "reward_std": 0.4658879339694977, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4165624976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7999999523162842, "step": 1867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 305.125, "completions/mean_terminated_length": 305.125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.3736, "grad_norm": 0.22397993505001068, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21708940.0, "reward": 1.2869229316711426, "reward_std": 0.39440444111824036, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.761343777179718, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 1868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 363.125, "completions/mean_terminated_length": 363.125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.3738, "grad_norm": 0.12858982384204865, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21718190.0, "reward": 0.7801569700241089, "reward_std": 0.2640874683856964, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4978124797344208, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 1869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 348.75, "completions/mean_terminated_length": 348.75, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.374, "grad_norm": 0.10135146230459213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21727770.0, "reward": 0.6965357661247253, "reward_std": 0.3984834849834442, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20556874573230743, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 1870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 474.4375, "completions/mean_terminated_length": 437.8000183105469, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.3742, "grad_norm": 0.14738987386226654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21740017.0, "reward": 0.8018510341644287, "reward_std": 0.5034181475639343, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3536999821662903, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.908750057220459, "step": 1871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 368.8125, "completions/mean_terminated_length": 368.8125, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.3744, "grad_norm": 0.010941114276647568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21750638.0, "reward": 0.5408479571342468, "reward_std": 0.01749608851969242, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08169375360012054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 495.6875, "completions/mean_terminated_length": 495.6875, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.3746, "grad_norm": 0.10705908387899399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21762537.0, "reward": 0.7579697966575623, "reward_std": 0.3808693587779999, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3284375071525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 1873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 527.4375, "completions/mean_terminated_length": 527.4375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.3748, "grad_norm": 0.017133206129074097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21775072.0, "reward": 0.5694509744644165, "reward_std": 0.08475174754858017, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1388999968767166, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 1874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 499.75, "completions/mean_terminated_length": 499.75, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.375, "grad_norm": 0.03973446041345596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21787260.0, "reward": 0.5924758911132812, "reward_std": 0.13253958523273468, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18494999408721924, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8787499666213989, "step": 1875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 274.625, "completions/mean_terminated_length": 274.625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.3752, "grad_norm": 0.12392152845859528, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21795302.0, "reward": 0.6870728135108948, "reward_std": 0.3373684287071228, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24914373457431793, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 1876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 271.875, "completions/mean_terminated_length": 271.875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.3754, "grad_norm": 0.1748495101928711, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21803540.0, "reward": 0.7189791202545166, "reward_std": 0.38723334670066833, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2504562437534332, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 651.4375, "completions/mean_terminated_length": 598.2142944335938, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.3756, "grad_norm": 0.033418480306863785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21818315.0, "reward": 0.5062384605407715, "reward_std": 0.22903187572956085, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13747499883174896, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7737500071525574, "step": 1878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 390.625, "completions/mean_terminated_length": 348.4000244140625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.3758, "grad_norm": 0.2580660283565521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21828277.0, "reward": 1.2377195358276367, "reward_std": 0.4878154993057251, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7879375219345093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8824999928474426, "step": 1879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 635.4375, "completions/mean_terminated_length": 609.5333862304688, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.376, "grad_norm": 0.0827951654791832, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21842300.0, "reward": 0.729844331741333, "reward_std": 0.2832881808280945, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4596874713897705, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.574999988079071, "step": 1880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 454.6875, "completions/mean_terminated_length": 454.6875, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.3762, "grad_norm": 0.0001476877514505759, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21853751.0, "reward": 1.499647855758667, "reward_std": 0.0005365860997699201, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9992936849594116, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9805625081062317, "step": 1881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 369.125, "completions/mean_terminated_length": 369.125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.3764, "grad_norm": 0.17994514107704163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21863641.0, "reward": 1.3179543018341064, "reward_std": 0.38982969522476196, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8234062194824219, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 1882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 325.4375, "completions/mean_terminated_length": 325.4375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.3766, "grad_norm": 0.1253189593553543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21872256.0, "reward": 1.3780328035354614, "reward_std": 0.3333521783351898, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8810624480247498, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9937500357627869, "step": 1883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 522.375, "completions/mean_terminated_length": 522.375, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.3768, "grad_norm": 0.09656046330928802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21884678.0, "reward": 1.2553918361663818, "reward_std": 0.4360753297805786, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7607812881469727, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 1884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 386.75, "completions/mean_terminated_length": 386.75, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.377, "grad_norm": 0.10669748485088348, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21894706.0, "reward": 0.6645291447639465, "reward_std": 0.3261067867279053, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2040562480688095, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 386.1875, "completions/mean_terminated_length": 386.1875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.3772, "grad_norm": 0.23816032707691193, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21904533.0, "reward": 0.9922665357589722, "reward_std": 0.473943829536438, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5470312833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 1886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 249.0, "completions/max_terminated_length": 249.0, "completions/mean_length": 196.125, "completions/mean_terminated_length": 196.125, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.3774, "grad_norm": 0.013507340103387833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21911575.0, "reward": 0.5244947671890259, "reward_std": 0.019738735631108284, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0489875003695488, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 1887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 596.125, "completions/mean_terminated_length": 567.6000366210938, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.3776, "grad_norm": 0.18782782554626465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21924937.0, "reward": 1.201432228088379, "reward_std": 0.49335533380508423, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7778624892234802, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8412500023841858, "step": 1888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 509.0, "completions/max_terminated_length": 509.0, "completions/mean_length": 443.0, "completions/mean_terminated_length": 443.0, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.3778, "grad_norm": 7.754137914162129e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21935641.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 676.875, "completions/mean_terminated_length": 653.7333374023438, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.378, "grad_norm": 0.09487307816743851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21950535.0, "reward": 1.2842572927474976, "reward_std": 0.475420206785202, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.818512499332428, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 1890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 980.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 747.25, "completions/mean_terminated_length": 747.25, "completions/min_length": 586.0, "completions/min_terminated_length": 586.0, "epoch": 0.3782, "grad_norm": 0.0843258947134018, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21966907.0, "reward": 1.131450891494751, "reward_std": 0.44402313232421875, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7003999948501587, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8862500190734863, "step": 1891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 331.875, "completions/mean_terminated_length": 331.875, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.3784, "grad_norm": 0.11686787754297256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21975945.0, "reward": 0.874453604221344, "reward_std": 0.197637677192688, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7489062547683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.34062498807907104, "step": 1892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 578.0625, "completions/mean_terminated_length": 578.0625, "completions/min_length": 498.0, "completions/min_terminated_length": 498.0, "epoch": 0.3786, "grad_norm": 0.055501341819763184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21989226.0, "reward": 1.4064167737960815, "reward_std": 0.269577294588089, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9378312826156616, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 1893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 924.3125, "completions/mean_terminated_length": 864.5, "completions/min_length": 723.0, "completions/min_terminated_length": 723.0, "epoch": 0.3788, "grad_norm": 0.10929036885499954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22008175.0, "reward": 0.4478224515914917, "reward_std": 0.47999870777130127, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14564375579357147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6081250309944153, "step": 1894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 436.0, "completions/max_terminated_length": 436.0, "completions/mean_length": 349.875, "completions/mean_terminated_length": 349.875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.379, "grad_norm": 0.3833540678024292, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22017725.0, "reward": 1.064213514328003, "reward_std": 0.6001125574111938, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6284250020980835, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.984375, "step": 1895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 715.625, "completions/mean_terminated_length": 695.0667114257812, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.3792, "grad_norm": 0.10607431828975677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22033031.0, "reward": 0.9726569652557373, "reward_std": 0.4671756327152252, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5703125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 1896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 315.0, "completions/max_terminated_length": 315.0, "completions/mean_length": 235.6875, "completions/mean_terminated_length": 235.6875, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.3794, "grad_norm": 0.09839450567960739, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22040450.0, "reward": 1.437779188156128, "reward_std": 0.24742142856121063, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.938056230545044, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 1897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 680.3125, "completions/mean_terminated_length": 657.4000244140625, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.3796, "grad_norm": 0.13690194487571716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22055271.0, "reward": 0.7278131246566772, "reward_std": 0.36875054240226746, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5181249976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6062500476837158, "step": 1898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 680.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 550.25, "completions/mean_terminated_length": 550.25, "completions/min_length": 456.0, "completions/min_terminated_length": 456.0, "epoch": 0.3798, "grad_norm": 0.11645545065402985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22068523.0, "reward": 1.0744822025299072, "reward_std": 0.49792665243148804, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5864624977111816, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 351.4375, "completions/mean_terminated_length": 351.4375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.38, "grad_norm": 0.15981894731521606, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22078114.0, "reward": 1.2536697387695312, "reward_std": 0.43906888365745544, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7573375105857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 1900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 960.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 569.75, "completions/mean_terminated_length": 569.75, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.3802, "grad_norm": 0.08079351484775543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22091438.0, "reward": 0.8758476972579956, "reward_std": 0.27155300974845886, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6891937255859375, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3956249952316284, "step": 1901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 527.875, "completions/mean_terminated_length": 527.875, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.3804, "grad_norm": 0.12566319108009338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22103708.0, "reward": 0.7735946178436279, "reward_std": 0.3748679757118225, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3596875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 1902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 379.4375, "completions/mean_terminated_length": 379.4375, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.3806, "grad_norm": 0.12825876474380493, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22114163.0, "reward": 0.8348323702812195, "reward_std": 0.46298637986183167, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3571624755859375, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 1903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 283.5625, "completions/mean_terminated_length": 283.5625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.3808, "grad_norm": 0.2944228947162628, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22122092.0, "reward": 1.2568416595458984, "reward_std": 0.4339075982570648, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7636812925338745, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 805.5625, "completions/mean_terminated_length": 755.1538696289062, "completions/min_length": 581.0, "completions/min_terminated_length": 581.0, "epoch": 0.381, "grad_norm": 0.06437337398529053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22139093.0, "reward": 0.4552350640296936, "reward_std": 0.2920733690261841, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16046875715255737, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.699999988079071, "step": 1905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 605.8125, "completions/mean_terminated_length": 605.8125, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.3812, "grad_norm": 0.1297626793384552, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22152706.0, "reward": 0.7509381771087646, "reward_std": 0.28132331371307373, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5643750429153442, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 1906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 949.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 575.6875, "completions/mean_terminated_length": 575.6875, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.3814, "grad_norm": 0.11303280293941498, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22165725.0, "reward": 0.9761635661125183, "reward_std": 0.43742167949676514, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5773249864578247, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8637499809265137, "step": 1907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 459.25, "completions/mean_terminated_length": 459.25, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.3816, "grad_norm": 0.022368403151631355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22176993.0, "reward": 0.5724196434020996, "reward_std": 0.09398841112852097, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14483749866485596, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 1908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 335.75, "completions/mean_terminated_length": 335.75, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.3818, "grad_norm": 0.16926607489585876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22186397.0, "reward": 1.4035695791244507, "reward_std": 0.20439882576465607, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9946374893188477, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7912499904632568, "step": 1909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 425.5, "completions/mean_terminated_length": 425.5, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.382, "grad_norm": 0.06170526146888733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22197285.0, "reward": 1.4351415634155273, "reward_std": 0.24683168530464172, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9327812790870667, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418750405311584, "step": 1910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 454.0, "completions/max_terminated_length": 454.0, "completions/mean_length": 343.625, "completions/mean_terminated_length": 343.625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.3822, "grad_norm": 0.09377136826515198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22207391.0, "reward": 1.4398728609085083, "reward_std": 0.23763319849967957, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422438144683838, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 1911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 305.8125, "completions/mean_terminated_length": 305.8125, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.3824, "grad_norm": 0.07125259935855865, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22216140.0, "reward": 1.439600944519043, "reward_std": 0.2375604808330536, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416999816894531, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 1912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 652.25, "completions/mean_terminated_length": 627.4666748046875, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.3826, "grad_norm": 0.11186046898365021, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22231296.0, "reward": 0.6106259226799011, "reward_std": 0.2712569832801819, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1587499976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 1913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 253.1875, "completions/mean_terminated_length": 253.1875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.3828, "grad_norm": 0.24721291661262512, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22238915.0, "reward": 0.9554321765899658, "reward_std": 0.4953717589378357, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4733625054359436, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 1914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 899.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 525.9375, "completions/mean_terminated_length": 525.9375, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.383, "grad_norm": 0.21865251660346985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22250978.0, "reward": 0.7850790619850159, "reward_std": 0.3845260441303253, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3826562166213989, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.840624988079071, "step": 1915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 241.0, "completions/max_terminated_length": 241.0, "completions/mean_length": 191.375, "completions/mean_terminated_length": 191.375, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 0.3832, "grad_norm": 0.2761673927307129, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22258168.0, "reward": 1.3195761442184448, "reward_std": 0.38727253675460815, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8266500234603882, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 1916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 413.9375, "completions/mean_terminated_length": 413.9375, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.3834, "grad_norm": 0.11254821717739105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22269703.0, "reward": 1.318200945854187, "reward_std": 0.3882570266723633, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8238999843597412, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 1917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 656.25, "completions/mean_terminated_length": 631.7333374023438, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.3836, "grad_norm": 0.04528307169675827, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22284395.0, "reward": 0.5741415023803711, "reward_std": 0.20334593951702118, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2107812464237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 1918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 973.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 700.75, "completions/mean_terminated_length": 700.75, "completions/min_length": 542.0, "completions/min_terminated_length": 542.0, "epoch": 0.3838, "grad_norm": 0.17775197327136993, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22303783.0, "reward": 0.964760422706604, "reward_std": 0.4870947599411011, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49201875925064087, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 345.0, "completions/max_terminated_length": 345.0, "completions/mean_length": 268.5, "completions/mean_terminated_length": 268.5, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.384, "grad_norm": 0.28827089071273804, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22312255.0, "reward": 1.2522671222686768, "reward_std": 0.4414333701133728, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7545312643051147, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9718749523162842, "step": 1920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 591.0625, "completions/mean_terminated_length": 591.0625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.3842, "grad_norm": 0.06661607325077057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22326592.0, "reward": 0.6323449611663818, "reward_std": 0.16873422265052795, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2646874785423279, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 1921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 864.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 449.75, "completions/mean_terminated_length": 449.75, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.3844, "grad_norm": 0.15232568979263306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22338028.0, "reward": 0.8013229370117188, "reward_std": 0.424601674079895, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3526437282562256, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331249594688416, "step": 1922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 612.9375, "completions/mean_terminated_length": 554.2142944335938, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.3846, "grad_norm": 0.13652828335762024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22351643.0, "reward": 1.0115509033203125, "reward_std": 0.594541072845459, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5856000185012817, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8362500071525574, "step": 1923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 326.3125, "completions/mean_terminated_length": 326.3125, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.3848, "grad_norm": 0.1195007786154747, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22361104.0, "reward": 0.5988293290138245, "reward_std": 0.24138756096363068, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13515624403953552, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 986.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 408.0625, "completions/mean_terminated_length": 408.0625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.385, "grad_norm": 0.2039504498243332, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22371521.0, "reward": 1.320857286453247, "reward_std": 0.3830723762512207, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829212486743927, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 1925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 381.1875, "completions/mean_terminated_length": 381.1875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.3852, "grad_norm": 0.26650136709213257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22381620.0, "reward": 0.9537041187286377, "reward_std": 0.4971279799938202, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4699062705039978, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 1926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 781.8125, "completions/mean_terminated_length": 725.923095703125, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.3854, "grad_norm": 0.13898658752441406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22397793.0, "reward": 1.12443208694458, "reward_std": 0.6704818606376648, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7488625049591064, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7868750095367432, "step": 1927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 293.375, "completions/mean_terminated_length": 293.375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.3856, "grad_norm": 0.11485537886619568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22406503.0, "reward": 1.4398698806762695, "reward_std": 0.2376323789358139, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422374963760376, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.966249942779541, "step": 1928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 560.4375, "completions/mean_terminated_length": 529.5333862304688, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.3858, "grad_norm": 0.051027584820985794, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22419006.0, "reward": 0.5264729857444763, "reward_std": 0.16694697737693787, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11544374376535416, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 1929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 243.0, "completions/max_terminated_length": 243.0, "completions/mean_length": 199.6875, "completions/mean_terminated_length": 199.6875, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.386, "grad_norm": 0.31387144327163696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22425833.0, "reward": 1.313010573387146, "reward_std": 0.4017254114151001, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.813518762588501, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9893749952316284, "step": 1930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 287.1875, "completions/mean_terminated_length": 287.1875, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.3862, "grad_norm": 0.0003633153100963682, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22434204.0, "reward": 1.4992759227752686, "reward_std": 0.0005422153044492006, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985499978065491, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 826.0, "completions/max_terminated_length": 826.0, "completions/mean_length": 696.5625, "completions/mean_terminated_length": 696.5625, "completions/min_length": 521.0, "completions/min_terminated_length": 521.0, "epoch": 0.3864, "grad_norm": 0.011183607392013073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22449349.0, "reward": 0.5625666379928589, "reward_std": 0.05596289411187172, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1251312494277954, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 1932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 991.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 719.75, "completions/mean_terminated_length": 719.75, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.3866, "grad_norm": 0.03795257583260536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22465393.0, "reward": 0.8592191934585571, "reward_std": 0.15817543864250183, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7184375524520874, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.45624998211860657, "step": 1933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 708.125, "completions/mean_terminated_length": 708.125, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.3868, "grad_norm": 0.0271566491574049, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22480355.0, "reward": 0.6704695820808411, "reward_std": 0.11229585856199265, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3409374952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7999999523162842, "step": 1934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 412.6875, "completions/mean_terminated_length": 412.6875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.387, "grad_norm": 0.27565672993659973, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22491950.0, "reward": 1.0156856775283813, "reward_std": 0.500106692314148, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5313687324523926, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 1935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 542.125, "completions/mean_terminated_length": 510.0000305175781, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.3872, "grad_norm": 0.14984863996505737, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22505024.0, "reward": 1.0058320760726929, "reward_std": 0.5063098073005676, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.636662483215332, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 1936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 390.0, "completions/mean_terminated_length": 390.0, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.3874, "grad_norm": 0.11808552592992783, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22515472.0, "reward": 1.2515323162078857, "reward_std": 0.44368264079093933, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7530624866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "step": 1937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 299.9375, "completions/mean_terminated_length": 299.9375, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.3876, "grad_norm": 0.22228462994098663, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22524031.0, "reward": 1.3771827220916748, "reward_std": 0.3330245018005371, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8793625235557556, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 1938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 505.125, "completions/mean_terminated_length": 505.125, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.3878, "grad_norm": 0.1300184726715088, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22536561.0, "reward": 0.7832041382789612, "reward_std": 0.42683979868888855, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 1939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 489.9375, "completions/mean_terminated_length": 454.3333435058594, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.388, "grad_norm": 0.07019639015197754, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22548256.0, "reward": 0.7925004959106445, "reward_std": 0.25817304849624634, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6475000381469727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4375, "step": 1940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 732.0, "completions/mean_length": 661.4375, "completions/mean_terminated_length": 577.7692260742188, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.3882, "grad_norm": 0.16241826117038727, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22562951.0, "reward": 0.7380319833755493, "reward_std": 0.5385485887527466, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.41356250643730164, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6575000286102295, "step": 1941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 810.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 582.875, "completions/mean_terminated_length": 582.875, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.3884, "grad_norm": 0.04775261506438255, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22576789.0, "reward": 0.6235164403915405, "reward_std": 0.14700189232826233, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24703124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8406249284744263, "step": 1942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 398.4375, "completions/mean_terminated_length": 398.4375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.3886, "grad_norm": 0.00018039013957604766, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22587452.0, "reward": 1.4990884065628052, "reward_std": 0.0005181925371289253, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981750249862671, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 1943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 500.5625, "completions/mean_terminated_length": 500.5625, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.3888, "grad_norm": 0.14346164464950562, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22599525.0, "reward": 1.2668633460998535, "reward_std": 0.414583683013916, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7837250232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 1944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 706.9375, "completions/mean_terminated_length": 685.800048828125, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.389, "grad_norm": 0.05926046893000603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22614644.0, "reward": 0.6099226474761963, "reward_std": 0.21359935402870178, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2823437452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7593749761581421, "step": 1945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 404.75, "completions/mean_terminated_length": 404.75, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.3892, "grad_norm": 0.07217945903539658, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22625248.0, "reward": 1.4398729801177979, "reward_std": 0.23763324320316315, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422436952590942, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 1946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 310.0625, "completions/mean_terminated_length": 310.0625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.3894, "grad_norm": 0.00022411427926272154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22634449.0, "reward": 1.4991728067398071, "reward_std": 0.0005645065102726221, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983437061309814, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 1947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 423.4375, "completions/mean_terminated_length": 423.4375, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.3896, "grad_norm": 0.15013116598129272, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22645896.0, "reward": 0.848516583442688, "reward_std": 0.4529082179069519, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38453125953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 1948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 228.5, "completions/mean_terminated_length": 228.5, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.3898, "grad_norm": 0.21512222290039062, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22653552.0, "reward": 0.7742009162902832, "reward_std": 0.43255558609962463, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29839998483657837, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 707.8125, "completions/mean_terminated_length": 707.8125, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.39, "grad_norm": 0.040282055735588074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22669133.0, "reward": 0.8123444318771362, "reward_std": 0.19089606404304504, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6246874928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5, "step": 1950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 386.625, "completions/mean_terminated_length": 386.625, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.3902, "grad_norm": 0.16219860315322876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22679815.0, "reward": 1.3784854412078857, "reward_std": 0.329588383436203, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8819687366485596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 1951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 402.4375, "completions/mean_terminated_length": 402.4375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.3904, "grad_norm": 0.00031622996903024614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22690206.0, "reward": 1.4995200634002686, "reward_std": 0.0006156161543913186, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.999037504196167, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 1952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 598.625, "completions/mean_terminated_length": 570.2667236328125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.3906, "grad_norm": 0.07370875775814056, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22703672.0, "reward": 0.6574759483337402, "reward_std": 0.36833465099334717, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25244998931884766, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8600000143051147, "step": 1953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 454.3125, "completions/mean_terminated_length": 454.3125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.3908, "grad_norm": 0.1532241851091385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22715165.0, "reward": 1.2879695892333984, "reward_std": 0.38865208625793457, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8259375095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 1954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 874.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 644.6875, "completions/mean_terminated_length": 644.6875, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.391, "grad_norm": 0.12162226438522339, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22730472.0, "reward": 1.0061728954315186, "reward_std": 0.4651046395301819, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5748437643051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 1955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 264.0, "completions/max_terminated_length": 264.0, "completions/mean_length": 220.375, "completions/mean_terminated_length": 220.375, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.3912, "grad_norm": 0.271224707365036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22737390.0, "reward": 1.1941510438919067, "reward_std": 0.46834123134613037, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7007999420166016, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9836875200271606, "step": 1956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 907.0625, "completions/mean_terminated_length": 853.9091186523438, "completions/min_length": 615.0, "completions/min_terminated_length": 615.0, "epoch": 0.3914, "grad_norm": 0.09186653047800064, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22756399.0, "reward": 0.6475005745887756, "reward_std": 0.5818286538124084, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35750001668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.612500011920929, "step": 1957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 329.5625, "completions/mean_terminated_length": 329.5625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.3916, "grad_norm": 0.17603006958961487, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22766232.0, "reward": 0.7802354693412781, "reward_std": 0.4289342761039734, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31046876311302185, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 1958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 475.0, "completions/max_terminated_length": 475.0, "completions/mean_length": 309.9375, "completions/mean_terminated_length": 309.9375, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.3918, "grad_norm": 0.26909327507019043, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22775335.0, "reward": 1.2552165985107422, "reward_std": 0.3517650365829468, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8854312896728516, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7256250381469727, "step": 1959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 506.75, "completions/mean_terminated_length": 506.75, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.392, "grad_norm": 0.12417672574520111, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22787795.0, "reward": 0.6892979145050049, "reward_std": 0.33594006299972534, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2535937428474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 1960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 394.375, "completions/mean_terminated_length": 394.375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.3922, "grad_norm": 0.14812229573726654, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22797945.0, "reward": 1.1402071714401245, "reward_std": 0.4784258306026459, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6554125547409058, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 1961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 756.8125, "completions/mean_terminated_length": 718.6428833007812, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.3924, "grad_norm": 0.15470293164253235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22813862.0, "reward": 1.1337382793426514, "reward_std": 0.5806760191917419, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7049750089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8350000381469727, "step": 1962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 428.25, "completions/mean_terminated_length": 428.25, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.3926, "grad_norm": 0.1703294813632965, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22825178.0, "reward": 1.320822834968567, "reward_std": 0.3830551505088806, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829143762588501, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 1963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 894.4375, "completions/mean_terminated_length": 835.5454711914062, "completions/min_length": 685.0, "completions/min_terminated_length": 685.0, "epoch": 0.3928, "grad_norm": 0.1324937343597412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22843649.0, "reward": 0.7987506985664368, "reward_std": 0.6747738718986511, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4724999964237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6499999761581421, "step": 1964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 305.75, "completions/mean_terminated_length": 305.75, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.393, "grad_norm": 0.21116550266742706, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22852253.0, "reward": 0.7711541652679443, "reward_std": 0.43441715836524963, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29230624437332153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 1965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 434.4375, "completions/mean_terminated_length": 434.4375, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.3932, "grad_norm": 0.10759596526622772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22863236.0, "reward": 0.7268759608268738, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26624998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 556.4375, "completions/mean_terminated_length": 556.4375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.3934, "grad_norm": 0.12630726397037506, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22876459.0, "reward": 1.261763572692871, "reward_std": 0.42515888810157776, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7735249996185303, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 1967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 600.75, "completions/mean_terminated_length": 600.75, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.3936, "grad_norm": 0.06293632090091705, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22892183.0, "reward": 0.6732822060585022, "reward_std": 0.3226090967655182, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2215624898672104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 1968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 533.4375, "completions/mean_terminated_length": 500.7333679199219, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.3938, "grad_norm": 0.14843228459358215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22904558.0, "reward": 1.0491291284561157, "reward_std": 0.5432435870170593, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5982562303543091, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 1969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 610.375, "completions/mean_terminated_length": 582.800048828125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.394, "grad_norm": 0.1058964654803276, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22917972.0, "reward": 0.63187575340271, "reward_std": 0.24240192770957947, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32625001668930054, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.706250011920929, "step": 1970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 516.875, "completions/mean_terminated_length": 516.875, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.3942, "grad_norm": 0.07045827805995941, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22930930.0, "reward": 0.6081259250640869, "reward_std": 0.23749999701976776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1537500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 465.9375, "completions/mean_terminated_length": 465.9375, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.3944, "grad_norm": 0.15411369502544403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22942929.0, "reward": 1.1705446243286133, "reward_std": 0.4069973826408386, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7785875201225281, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8025000095367432, "step": 1972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 601.375, "completions/mean_terminated_length": 601.375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.3946, "grad_norm": 0.11131642013788223, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22957351.0, "reward": 0.9688289165496826, "reward_std": 0.4351949989795685, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5626562833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 1973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 463.625, "completions/mean_terminated_length": 276.8333435058594, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.3948, "grad_norm": 0.31427261233329773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22968433.0, "reward": 0.7655320167541504, "reward_std": 0.6257521510124207, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4060624837875366, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7200000286102295, "step": 1974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 508.8125, "completions/mean_terminated_length": 508.8125, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.395, "grad_norm": 0.10633537173271179, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22980302.0, "reward": 1.320847988128662, "reward_std": 0.38306766748428345, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291937112808228, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 457.625, "completions/mean_terminated_length": 457.625, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.3952, "grad_norm": 0.00014960776024963707, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22991256.0, "reward": 1.499091625213623, "reward_std": 0.0005239492165856063, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981812238693237, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 1976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 477.125, "completions/mean_terminated_length": 477.125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.3954, "grad_norm": 0.15988770127296448, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23002890.0, "reward": 1.0692946910858154, "reward_std": 0.5289546251296997, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.638587474822998, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9243749976158142, "step": 1977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 560.1875, "completions/mean_terminated_length": 560.1875, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.3956, "grad_norm": 0.05179842561483383, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23015853.0, "reward": 0.6008447408676147, "reward_std": 0.23997217416763306, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1391874998807907, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 1978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 750.625, "completions/mean_terminated_length": 732.4000244140625, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.3958, "grad_norm": 0.04230485484004021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23032407.0, "reward": 0.7282818555831909, "reward_std": 0.2352285087108612, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5190625190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.59375, "step": 1979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 517.0, "completions/mean_terminated_length": 517.0, "completions/min_length": 463.0, "completions/min_terminated_length": 463.0, "epoch": 0.396, "grad_norm": 0.07946556806564331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23044263.0, "reward": 1.4373698234558105, "reward_std": 0.24731232225894928, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9372375011444092, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 1980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 386.0, "completions/mean_terminated_length": 386.0, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.3962, "grad_norm": 0.05334876850247383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23054647.0, "reward": 0.6063165664672852, "reward_std": 0.23809203505516052, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1501312404870987, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 637.4375, "completions/mean_terminated_length": 611.6666870117188, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.3964, "grad_norm": 0.07951679825782776, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23070478.0, "reward": 0.7207040190696716, "reward_std": 0.3950577676296234, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 1982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 347.3125, "completions/mean_terminated_length": 347.3125, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.3966, "grad_norm": 0.14903666079044342, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23079827.0, "reward": 0.5961134433746338, "reward_std": 0.24155794084072113, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12972497940063477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 1983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 795.5, "completions/mean_terminated_length": 762.857177734375, "completions/min_length": 504.0, "completions/min_terminated_length": 504.0, "epoch": 0.3968, "grad_norm": 0.09789538383483887, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23096699.0, "reward": 0.8720163702964783, "reward_std": 0.5504990816116333, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49403122067451477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7793749570846558, "step": 1984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 411.875, "completions/mean_terminated_length": 411.875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.397, "grad_norm": 0.18164537847042084, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23107241.0, "reward": 0.8414697647094727, "reward_std": 0.45798641443252563, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37043750286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 1985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 704.125, "completions/mean_terminated_length": 658.4285888671875, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.3972, "grad_norm": 0.11779651790857315, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23122555.0, "reward": 0.8854479193687439, "reward_std": 0.5838289856910706, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45839375257492065, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.846875011920929, "step": 1986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 532.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 366.625, "completions/mean_terminated_length": 366.625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.3974, "grad_norm": 0.11127305775880814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23133333.0, "reward": 1.3780728578567505, "reward_std": 0.33138328790664673, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8811437487602234, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 1987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 532.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 440.25, "completions/mean_terminated_length": 440.25, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.3976, "grad_norm": 0.21606069803237915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23144121.0, "reward": 0.9637010097503662, "reward_std": 0.4902438521385193, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48989999294281006, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562500715255737, "step": 1988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 664.0, "completions/max_terminated_length": 664.0, "completions/mean_length": 486.625, "completions/mean_terminated_length": 486.625, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.3978, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23156099.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 409.25, "completions/mean_terminated_length": 409.25, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.398, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23166439.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 1990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 386.4375, "completions/mean_terminated_length": 386.4375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.3982, "grad_norm": 0.18207649886608124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23176462.0, "reward": 0.9613947868347168, "reward_std": 0.4897260367870331, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48528751730918884, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 1991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 516.6875, "completions/mean_terminated_length": 516.6875, "completions/min_length": 464.0, "completions/min_terminated_length": 464.0, "epoch": 0.3984, "grad_norm": 0.12579593062400818, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23188713.0, "reward": 0.962566614151001, "reward_std": 0.48842906951904297, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48763126134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 1992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 583.375, "completions/mean_terminated_length": 583.375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.3986, "grad_norm": 0.10534843802452087, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23202671.0, "reward": 0.9228134155273438, "reward_std": 0.3051066994667053, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.7206250429153442, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4437500238418579, "step": 1993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 490.9375, "completions/mean_terminated_length": 490.9375, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.3988, "grad_norm": 0.016271384432911873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23214446.0, "reward": 0.5598321557044983, "reward_std": 0.05254356563091278, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11966249346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 1994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 258.5625, "completions/mean_terminated_length": 258.5625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.399, "grad_norm": 0.15625838935375214, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23222823.0, "reward": 1.380285382270813, "reward_std": 0.32459819316864014, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855687379837036, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 1995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 822.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 680.5625, "completions/mean_terminated_length": 680.5625, "completions/min_length": 493.0, "completions/min_terminated_length": 493.0, "epoch": 0.3992, "grad_norm": 0.07641954720020294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23237472.0, "reward": 0.7785042524337769, "reward_std": 0.4301752746105194, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30700626969337463, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 1996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 584.25, "completions/mean_terminated_length": 554.933349609375, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.3994, "grad_norm": 0.14283128082752228, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23250996.0, "reward": 1.044191598892212, "reward_std": 0.5480945110321045, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5883812308311462, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9031250476837158, "step": 1997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 319.1875, "completions/mean_terminated_length": 319.1875, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.3996, "grad_norm": 0.2101457566022873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23259463.0, "reward": 0.778397798538208, "reward_std": 0.4297454059123993, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.306793749332428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 1998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 681.0625, "completions/mean_terminated_length": 632.0714721679688, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.3998, "grad_norm": 0.2046000063419342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23274376.0, "reward": 0.8105477094650269, "reward_std": 0.5307909846305847, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.43359375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 1999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 627.375, "completions/mean_terminated_length": 570.7142944335938, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.4, "grad_norm": 0.06821789592504501, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23288654.0, "reward": 0.49514463543891907, "reward_std": 0.20578418672084808, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11528749763965607, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8112499713897705, "step": 2000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 621.5, "completions/mean_terminated_length": 564.0, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.4002, "grad_norm": 0.06906577944755554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23303270.0, "reward": 0.5546882748603821, "reward_std": 0.26671287417411804, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.234375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7062499523162842, "step": 2001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 629.375, "completions/mean_terminated_length": 603.0667114257812, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.4004, "grad_norm": 0.10324966907501221, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23318460.0, "reward": 0.6161668300628662, "reward_std": 0.3693169951438904, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16983124613761902, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 2002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 407.6875, "completions/mean_terminated_length": 407.6875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.4006, "grad_norm": 0.16052845120429993, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23328711.0, "reward": 0.9577165246009827, "reward_std": 0.4929707944393158, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47793126106262207, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 378.375, "completions/mean_terminated_length": 378.375, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.4008, "grad_norm": 0.2876659035682678, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23339917.0, "reward": 0.9638165235519409, "reward_std": 0.48608121275901794, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4901312589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456249475479126, "step": 2004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 396.25, "completions/mean_terminated_length": 354.4000244140625, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.401, "grad_norm": 0.1864277422428131, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23350001.0, "reward": 1.2923258543014526, "reward_std": 0.4617362320423126, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8346499800682068, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8912500143051147, "step": 2005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 321.75, "completions/mean_terminated_length": 321.75, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.4012, "grad_norm": 0.34865859150886536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23358925.0, "reward": 1.0790446996688843, "reward_std": 0.49194303154945374, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5955874919891357, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 2006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 797.4375, "completions/mean_terminated_length": 765.0714721679688, "completions/min_length": 487.0, "completions/min_terminated_length": 487.0, "epoch": 0.4014, "grad_norm": 0.0895966961979866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23375684.0, "reward": 0.6578132510185242, "reward_std": 0.4001385569572449, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.31562501192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 2007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 278.0, "completions/mean_terminated_length": 278.0, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.4016, "grad_norm": 0.08437487483024597, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23383572.0, "reward": 1.4398291110992432, "reward_std": 0.23762160539627075, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.942156195640564, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 2008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 573.5625, "completions/mean_terminated_length": 573.5625, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.4018, "grad_norm": 0.02078607678413391, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23396557.0, "reward": 0.5912508964538574, "reward_std": 0.09849478304386139, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18250000476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.893750011920929, "step": 2009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 310.6875, "completions/mean_terminated_length": 310.6875, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.402, "grad_norm": 0.1314762681722641, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23405208.0, "reward": 1.3828916549682617, "reward_std": 0.31670066714286804, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8907812237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 373.0, "completions/max_terminated_length": 373.0, "completions/mean_length": 298.875, "completions/mean_terminated_length": 298.875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.4022, "grad_norm": 0.20510520040988922, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23413670.0, "reward": 1.4378384351730347, "reward_std": 0.24481046199798584, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9381749629974365, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 2011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 522.8125, "completions/mean_terminated_length": 489.4000244140625, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.4024, "grad_norm": 0.023339567705988884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23426851.0, "reward": 0.49245429039001465, "reward_std": 0.13393881916999817, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04740624874830246, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9131250381469727, "step": 2012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 509.1875, "completions/mean_terminated_length": 509.1875, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.4026, "grad_norm": 0.07784754037857056, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23438790.0, "reward": 1.4393761157989502, "reward_std": 0.23750267922878265, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 917.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 581.0625, "completions/mean_terminated_length": 581.0625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.4028, "grad_norm": 0.09396913647651672, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23452439.0, "reward": 1.2201571464538574, "reward_std": 0.3574879765510559, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.8778125047683716, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6749999523162842, "step": 2014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 732.875, "completions/mean_terminated_length": 713.4666748046875, "completions/min_length": 456.0, "completions/min_terminated_length": 456.0, "epoch": 0.403, "grad_norm": 0.09004753828048706, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23468421.0, "reward": 0.7519537806510925, "reward_std": 0.3332115411758423, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.50390625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5843750238418579, "step": 2015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 782.5, "completions/mean_terminated_length": 672.727294921875, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.4032, "grad_norm": 0.1440187394618988, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23485069.0, "reward": 0.7485756874084473, "reward_std": 0.6530113220214844, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4346500039100647, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6349999904632568, "step": 2016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 861.375, "completions/mean_terminated_length": 807.1666870117188, "completions/min_length": 612.0, "completions/min_terminated_length": 612.0, "epoch": 0.4034, "grad_norm": 0.09910890460014343, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23503267.0, "reward": 0.666485071182251, "reward_std": 0.4980085492134094, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.45796874165534973, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.4593749940395355, "step": 2017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 486.3125, "completions/mean_terminated_length": 486.3125, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.4036, "grad_norm": 0.1974848061800003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23514936.0, "reward": 0.7336729764938354, "reward_std": 0.3836364150047302, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2798437476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 316.5, "completions/mean_terminated_length": 316.5, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.4038, "grad_norm": 0.16826659440994263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23523696.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 569.125, "completions/mean_terminated_length": 569.125, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.404, "grad_norm": 0.09698337316513062, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23536626.0, "reward": 0.6572978496551514, "reward_std": 0.3290681838989258, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18959374725818634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 738.25, "completions/mean_terminated_length": 566.7999877929688, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.4042, "grad_norm": 0.1351209431886673, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23552198.0, "reward": 0.5008601546287537, "reward_std": 0.4758884012699127, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.31421875953674316, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.40312498807907104, "step": 2021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 629.625, "completions/mean_terminated_length": 629.625, "completions/min_length": 469.0, "completions/min_terminated_length": 469.0, "epoch": 0.4044, "grad_norm": 0.04056638479232788, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23566208.0, "reward": 1.439507246017456, "reward_std": 0.2375352531671524, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 733.875, "completions/mean_terminated_length": 602.0, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.4046, "grad_norm": 0.13593731820583344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23581790.0, "reward": 0.4959380030632019, "reward_std": 0.37077799439430237, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3043749928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.48750001192092896, "step": 2023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 468.6875, "completions/mean_terminated_length": 468.6875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.4048, "grad_norm": 0.029335256665945053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23593833.0, "reward": 0.5898416042327881, "reward_std": 0.11373139917850494, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17968124151229858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8931249380111694, "step": 2024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 547.3125, "completions/mean_terminated_length": 515.5333862304688, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.405, "grad_norm": 0.10391949117183685, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23607102.0, "reward": 0.5895944833755493, "reward_std": 0.22816981375217438, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24168750643730164, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7637499570846558, "step": 2025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 540.1875, "completions/mean_terminated_length": 540.1875, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.4052, "grad_norm": 0.005584688391536474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23620193.0, "reward": 0.5244824886322021, "reward_std": 0.02039940096437931, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04896249994635582, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 430.0, "completions/mean_terminated_length": 390.4000244140625, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.4054, "grad_norm": 0.15829429030418396, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23631649.0, "reward": 0.8766258955001831, "reward_std": 0.5166980028152466, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44075000286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8887499570846558, "step": 2027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 408.25, "completions/mean_terminated_length": 408.25, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.4056, "grad_norm": 0.12539532780647278, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23642373.0, "reward": 1.3771729469299316, "reward_std": 0.33301693201065063, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8793437480926514, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 2028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 414.0625, "completions/mean_terminated_length": 414.0625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.4058, "grad_norm": 0.24533213675022125, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23652422.0, "reward": 1.2019416093826294, "reward_std": 0.45482364296913147, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.716381311416626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 429.9375, "completions/mean_terminated_length": 429.9375, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.406, "grad_norm": 0.16001246869564056, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23663333.0, "reward": 1.2303136587142944, "reward_std": 0.42472097277641296, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7731249928474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 2030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1004.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 650.6875, "completions/mean_terminated_length": 650.6875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.4062, "grad_norm": 0.04117445647716522, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23678256.0, "reward": 0.6615570187568665, "reward_std": 0.200786292552948, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32311248779296875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7299999594688416, "step": 2031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 252.3125, "completions/mean_terminated_length": 252.3125, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.4064, "grad_norm": 0.16091755032539368, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23686021.0, "reward": 1.376794695854187, "reward_std": 0.33453649282455444, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8785874843597412, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 2032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 310.875, "completions/mean_terminated_length": 310.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.4066, "grad_norm": 0.3074079155921936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23695315.0, "reward": 0.7750009298324585, "reward_std": 0.3828461468219757, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.36250001192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 2033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 350.5, "completions/mean_terminated_length": 350.5, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.4068, "grad_norm": 0.14668941497802734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23704587.0, "reward": 1.439741611480713, "reward_std": 0.23759816586971283, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419812560081482, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 348.375, "completions/mean_terminated_length": 348.375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.407, "grad_norm": 0.12906721234321594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23713921.0, "reward": 1.3769540786743164, "reward_std": 0.3329312205314636, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.87890625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 351.125, "completions/mean_terminated_length": 351.125, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.4072, "grad_norm": 0.2738061249256134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23723331.0, "reward": 1.258360505104065, "reward_std": 0.430534690618515, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7667187452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 2036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 817.5625, "completions/mean_terminated_length": 769.923095703125, "completions/min_length": 602.0, "completions/min_terminated_length": 602.0, "epoch": 0.4074, "grad_norm": 0.12742699682712555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23740748.0, "reward": 0.9856662154197693, "reward_std": 0.5747172236442566, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7213312387466431, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5681250095367432, "step": 2037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 654.25, "completions/mean_terminated_length": 531.0, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.4076, "grad_norm": 0.23713092505931854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23755312.0, "reward": 0.6110942959785461, "reward_std": 0.44260480999946594, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.40968748927116394, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.4312500059604645, "step": 2038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 500.5625, "completions/mean_terminated_length": 500.5625, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.4078, "grad_norm": 0.028346354141831398, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23767145.0, "reward": 0.6332820653915405, "reward_std": 0.1110440120100975, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26656249165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 2039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 543.0625, "completions/mean_terminated_length": 543.0625, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.408, "grad_norm": 0.10144894570112228, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23779594.0, "reward": 0.8117196559906006, "reward_std": 0.41532108187675476, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3734375238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 2040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 960.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 556.625, "completions/mean_terminated_length": 556.625, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.4082, "grad_norm": 0.16586075723171234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23792244.0, "reward": 1.0014071464538574, "reward_std": 0.45764726400375366, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5653125047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 2041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 788.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 546.5, "completions/mean_terminated_length": 546.5, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.4084, "grad_norm": 0.12510807812213898, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23805084.0, "reward": 1.0456256866455078, "reward_std": 0.35053354501724243, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.778749942779541, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.59375, "step": 2042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 956.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 473.8125, "completions/mean_terminated_length": 473.8125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.4086, "grad_norm": 0.09924839437007904, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23816345.0, "reward": 0.49312645196914673, "reward_std": 0.13373015820980072, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04874999821186066, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 2043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 239.9375, "completions/mean_terminated_length": 239.9375, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.4088, "grad_norm": 0.41650617122650146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23823816.0, "reward": 1.1950790882110596, "reward_std": 0.4663845896720886, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7026562690734863, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 2044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 584.75, "completions/mean_terminated_length": 555.4666748046875, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.409, "grad_norm": 0.07588768750429153, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23837956.0, "reward": 0.5678010582923889, "reward_std": 0.28338077664375305, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1356000006198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 2045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 513.6875, "completions/mean_terminated_length": 513.6875, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.4092, "grad_norm": 0.05663695186376572, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23850287.0, "reward": 0.6505476832389832, "reward_std": 0.2296123057603836, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3635937571525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.690625011920929, "step": 2046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 705.625, "completions/mean_terminated_length": 599.5, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.4094, "grad_norm": 0.12121804803609848, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23866457.0, "reward": 0.527332067489624, "reward_std": 0.44770097732543945, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17966249585151672, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7174999713897705, "step": 2047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 466.5, "completions/mean_terminated_length": 466.5, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.4096, "grad_norm": 0.1432264745235443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23877921.0, "reward": 1.2161729335784912, "reward_std": 0.4316643476486206, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7448437213897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 2048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 490.1875, "completions/mean_terminated_length": 454.60003662109375, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.4098, "grad_norm": 0.11236307770013809, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23889108.0, "reward": 0.7471634149551392, "reward_std": 0.4680781662464142, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30682501196861267, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 2049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 580.5625, "completions/mean_terminated_length": 478.23077392578125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.41, "grad_norm": 0.07945522665977478, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23902573.0, "reward": 0.5579694509506226, "reward_std": 0.3834210932254791, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2409374862909317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.668749988079071, "step": 2050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 826.0, "completions/mean_length": 619.5625, "completions/mean_terminated_length": 561.7857666015625, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.4102, "grad_norm": 0.11689579486846924, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23916070.0, "reward": 0.7527326345443726, "reward_std": 0.5459128022193909, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31796249747276306, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 2051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 424.0625, "completions/mean_terminated_length": 424.0625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.4104, "grad_norm": 0.21175003051757812, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23927287.0, "reward": 0.9432039856910706, "reward_std": 0.4413711726665497, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5114062428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8718749284744263, "step": 2052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 488.625, "completions/mean_terminated_length": 488.625, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.4106, "grad_norm": 0.14053307473659515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23938785.0, "reward": 0.6554479598999023, "reward_std": 0.32994556427001953, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18589374423027039, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 2053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 654.4375, "completions/mean_terminated_length": 601.6428833007812, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.4108, "grad_norm": 0.1391780525445938, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23953128.0, "reward": 0.8923853039741516, "reward_std": 0.5795878171920776, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4722687602043152, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.84312504529953, "step": 2054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 948.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 753.125, "completions/mean_terminated_length": 753.125, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.411, "grad_norm": 0.12030511349439621, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23969578.0, "reward": 0.9027071595191956, "reward_std": 0.4772097170352936, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43041250109672546, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 2055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 421.75, "completions/mean_terminated_length": 421.75, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.4112, "grad_norm": 0.16825784742832184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23981414.0, "reward": 1.2019541263580322, "reward_std": 0.45483243465423584, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7164062261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 468.5625, "completions/mean_terminated_length": 468.5625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.4114, "grad_norm": 0.14987722039222717, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23992479.0, "reward": 0.8207577466964722, "reward_std": 0.4725382924079895, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.32901251316070557, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 2057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 292.0625, "completions/mean_terminated_length": 292.0625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.4116, "grad_norm": 0.00022214205819182098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24000800.0, "reward": 1.4995572566986084, "reward_std": 0.0005636073183268309, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991124868392944, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "step": 2058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 741.375, "completions/mean_terminated_length": 701.0000610351562, "completions/min_length": 443.0, "completions/min_terminated_length": 443.0, "epoch": 0.4118, "grad_norm": 0.11338803917169571, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24017702.0, "reward": 0.7468756437301636, "reward_std": 0.4371790885925293, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4312500059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.675000011920929, "step": 2059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 685.625, "completions/mean_terminated_length": 637.2857666015625, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.412, "grad_norm": 0.11497373878955841, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24033152.0, "reward": 1.2144538164138794, "reward_std": 0.5145761966705322, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8664062023162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6593749523162842, "step": 2060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 335.875, "completions/mean_terminated_length": 335.875, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.4122, "grad_norm": 0.0619753934442997, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24042606.0, "reward": 0.5102196931838989, "reward_std": 0.13655661046504974, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08293750137090683, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 434.125, "completions/mean_terminated_length": 394.8000183105469, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.4124, "grad_norm": 0.06610318273305893, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24053568.0, "reward": 0.571403980255127, "reward_std": 0.28247976303100586, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428062617778778, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 2062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 457.1875, "completions/mean_terminated_length": 457.1875, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.4126, "grad_norm": 0.14131034910678864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24064691.0, "reward": 0.8456259369850159, "reward_std": 0.4547778367996216, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3787499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 374.8125, "completions/mean_terminated_length": 374.8125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.4128, "grad_norm": 0.10408460348844528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24074704.0, "reward": 0.6910821199417114, "reward_std": 0.33133524656295776, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25716251134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 2064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 300.0, "completions/max_terminated_length": 300.0, "completions/mean_length": 204.5, "completions/mean_terminated_length": 204.5, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 0.413, "grad_norm": 0.2793230712413788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24081496.0, "reward": 1.3193883895874023, "reward_std": 0.38717949390411377, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8262749910354614, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 826.0625, "completions/mean_terminated_length": 797.7857666015625, "completions/min_length": 617.0, "completions/min_terminated_length": 617.0, "epoch": 0.4132, "grad_norm": 0.06135662645101547, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24098905.0, "reward": 0.6432819366455078, "reward_std": 0.3644510805606842, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3490625023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.668749988079071, "step": 2066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 474.5625, "completions/mean_terminated_length": 474.5625, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.4134, "grad_norm": 0.13942226767539978, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24110530.0, "reward": 1.2595727443695068, "reward_std": 0.42821696400642395, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691437602043152, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 1024.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 899.4375, "completions/mean_terminated_length": 739.2857666015625, "completions/min_length": 550.0, "completions/min_terminated_length": 550.0, "epoch": 0.4136, "grad_norm": 0.1290547102689743, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24130009.0, "reward": 0.41820353269577026, "reward_std": 0.5856600999832153, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21140624582767487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.4156249761581421, "step": 2068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 384.0, "completions/mean_terminated_length": 384.0, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.4138, "grad_norm": 0.050895072519779205, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24140169.0, "reward": 0.5895321369171143, "reward_std": 0.1114354208111763, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17906250059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 2069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 269.625, "completions/mean_terminated_length": 269.625, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.414, "grad_norm": 0.17609399557113647, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24148275.0, "reward": 1.320766568183899, "reward_std": 0.38302719593048096, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829031229019165, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 418.0, "completions/mean_terminated_length": 418.0, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.4142, "grad_norm": 0.041517190635204315, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24158611.0, "reward": 0.9229691028594971, "reward_std": 0.1067509576678276, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8459375500679016, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3125, "step": 2071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 253.3125, "completions/mean_terminated_length": 253.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.4144, "grad_norm": 0.17924627661705017, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24166968.0, "reward": 1.3206915855407715, "reward_std": 0.38298991322517395, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8288812637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 617.6875, "completions/mean_terminated_length": 559.6428833007812, "completions/min_length": 475.0, "completions/min_terminated_length": 475.0, "epoch": 0.4146, "grad_norm": 0.11471456289291382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24181427.0, "reward": 1.217888355255127, "reward_std": 0.5462979078292847, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8107750415802002, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7774999737739563, "step": 2073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 768.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 544.1875, "completions/mean_terminated_length": 544.1875, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.4148, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24193958.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 488.8125, "completions/mean_terminated_length": 453.13336181640625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.415, "grad_norm": 0.08906810730695724, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24206259.0, "reward": 1.2893729209899902, "reward_std": 0.46648290753364563, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287437558174133, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 2075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 257.125, "completions/mean_terminated_length": 257.125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.4152, "grad_norm": 0.2116217315196991, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24213893.0, "reward": 1.3751978874206543, "reward_std": 0.3397049903869629, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8753937482833862, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 2076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 488.75, "completions/mean_terminated_length": 488.75, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.4154, "grad_norm": 0.07095184177160263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24225569.0, "reward": 0.6763696074485779, "reward_std": 0.28721654415130615, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2902374863624573, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8137499690055847, "step": 2077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 931.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 443.5, "completions/mean_terminated_length": 443.5, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4156, "grad_norm": 0.0027754344046115875, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24236297.0, "reward": 0.5469415187835693, "reward_std": 0.007237497717142105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09388124942779541, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 428.25, "completions/mean_terminated_length": 388.5333557128906, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.4158, "grad_norm": 0.1559438854455948, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24247085.0, "reward": 1.0457820892333984, "reward_std": 0.5460990071296692, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5915625095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 2079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 523.1875, "completions/mean_terminated_length": 489.8000183105469, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.416, "grad_norm": 0.1490137130022049, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24259296.0, "reward": 1.1675790548324585, "reward_std": 0.5229057669639587, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.710156261920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 297.4375, "completions/mean_terminated_length": 297.4375, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.4162, "grad_norm": 0.00033098121639341116, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24267927.0, "reward": 1.4993321895599365, "reward_std": 0.0006022312445566058, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998662531375885, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 564.75, "completions/mean_terminated_length": 564.75, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.4164, "grad_norm": 0.09036100655794144, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24280787.0, "reward": 1.2066946029663086, "reward_std": 0.4156629145145416, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7258875370025635, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8537499904632568, "step": 2082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 480.25, "completions/mean_terminated_length": 480.25, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.4166, "grad_norm": 0.1332612931728363, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24292151.0, "reward": 1.1972885131835938, "reward_std": 0.46255040168762207, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7070749998092651, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 603.4375, "completions/mean_terminated_length": 575.4000244140625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.4168, "grad_norm": 0.1956600844860077, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24305694.0, "reward": 0.9356885552406311, "reward_std": 0.5010125041007996, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5588747262954712, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7886250019073486, "step": 2084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 682.25, "completions/mean_terminated_length": 633.4285888671875, "completions/min_length": 449.0, "completions/min_terminated_length": 449.0, "epoch": 0.417, "grad_norm": 0.12020065635442734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24320866.0, "reward": 0.769219696521759, "reward_std": 0.4975036680698395, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.41343748569488525, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 2085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 384.0, "completions/mean_length": 391.375, "completions/mean_terminated_length": 349.20001220703125, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.4172, "grad_norm": 0.1710837334394455, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24330856.0, "reward": 1.311935305595398, "reward_std": 0.5121272802352905, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8738687634468079, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 2086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 490.875, "completions/mean_terminated_length": 490.875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.4174, "grad_norm": 0.13504652678966522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24342646.0, "reward": 1.095704197883606, "reward_std": 0.4739232361316681, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.62890625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 2087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 351.4375, "completions/mean_terminated_length": 351.4375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.4176, "grad_norm": 0.2431914061307907, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24352173.0, "reward": 0.7844415903091431, "reward_std": 0.4259919226169586, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31888124346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 415.4375, "completions/mean_terminated_length": 374.86669921875, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.4178, "grad_norm": 0.15099003911018372, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24362996.0, "reward": 0.8095196485519409, "reward_std": 0.49842825531959534, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3690375089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 2089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 597.0625, "completions/mean_terminated_length": 568.6000366210938, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.418, "grad_norm": 0.12052915245294571, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24376677.0, "reward": 0.7918758392333984, "reward_std": 0.460509717464447, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39625000953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8125, "step": 2090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 921.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 398.9375, "completions/mean_terminated_length": 398.9375, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.4182, "grad_norm": 0.10499512404203415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24387220.0, "reward": 0.6866414546966553, "reward_std": 0.26084932684898376, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.31078124046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "step": 2091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 445.625, "completions/mean_terminated_length": 445.625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.4184, "grad_norm": 0.09688303619623184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24398430.0, "reward": 1.2830355167388916, "reward_std": 0.40056225657463074, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8160687685012817, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081250429153442, "step": 2092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 664.0625, "completions/mean_terminated_length": 640.0667114257812, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.4186, "grad_norm": 0.12069923430681229, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24413631.0, "reward": 0.8528133630752563, "reward_std": 0.48326101899147034, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4556249976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 2093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 552.5, "completions/mean_terminated_length": 552.5, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.4188, "grad_norm": 0.02492854744195938, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24426519.0, "reward": 0.578035295009613, "reward_std": 0.11339680850505829, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15606874227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8893750309944153, "step": 2094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 539.5, "completions/mean_terminated_length": 539.5, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.419, "grad_norm": 0.08303002268075943, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24439151.0, "reward": 1.2206008434295654, "reward_std": 0.4333614110946655, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7537000179290771, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287500381469727, "step": 2095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 339.9375, "completions/mean_terminated_length": 339.9375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.4192, "grad_norm": 0.00019089228590019047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24448606.0, "reward": 1.4988821744918823, "reward_std": 0.0003586417296901345, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 324.375, "completions/mean_terminated_length": 324.375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.4194, "grad_norm": 0.13697639107704163, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24457908.0, "reward": 1.3183414936065674, "reward_std": 0.3883269727230072, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.824181318283081, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 384.25, "completions/mean_terminated_length": 384.25, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.4196, "grad_norm": 0.12715785205364227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24467736.0, "reward": 1.2615227699279785, "reward_std": 0.42501530051231384, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7730437517166138, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 465.25, "completions/mean_terminated_length": 465.25, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.4198, "grad_norm": 0.005541563034057617, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24479212.0, "reward": 0.5492165684700012, "reward_std": 0.015579955652356148, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09843124449253082, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 2099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 888.0, "completions/mean_terminated_length": 806.4000244140625, "completions/min_length": 580.0, "completions/min_terminated_length": 580.0, "epoch": 0.42, "grad_norm": 0.08576670289039612, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24497596.0, "reward": 0.41162869334220886, "reward_std": 0.4023786783218384, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.135756254196167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5831249952316284, "step": 2100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 521.1875, "completions/mean_terminated_length": 521.1875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.4202, "grad_norm": 0.11028574407100677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24510175.0, "reward": 1.081316590309143, "reward_std": 0.48889806866645813, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6001312732696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 324.625, "completions/mean_terminated_length": 324.625, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.4204, "grad_norm": 0.12907235324382782, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24519241.0, "reward": 1.3739855289459229, "reward_std": 0.3411683440208435, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8729687333106995, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 2102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 755.5, "completions/mean_terminated_length": 737.6000366210938, "completions/min_length": 544.0, "completions/min_terminated_length": 544.0, "epoch": 0.4206, "grad_norm": 0.07380841672420502, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24535649.0, "reward": 0.7082727551460266, "reward_std": 0.4254865050315857, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29154375195503235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8731250166893005, "step": 2103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 391.5, "completions/mean_terminated_length": 391.5, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.4208, "grad_norm": 0.10536430031061172, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24546009.0, "reward": 0.7862509489059448, "reward_std": 0.4248528778553009, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 373.125, "completions/mean_terminated_length": 373.125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.421, "grad_norm": 0.1551789492368698, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24556011.0, "reward": 1.320932388305664, "reward_std": 0.3831096291542053, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8293625116348267, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 2105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 302.875, "completions/mean_terminated_length": 302.875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.4212, "grad_norm": 0.28264158964157104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24564601.0, "reward": 1.2567698955535889, "reward_std": 0.43389490246772766, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7635375261306763, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 644.0625, "completions/mean_terminated_length": 644.0625, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.4214, "grad_norm": 0.15896376967430115, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24578778.0, "reward": 1.1052355766296387, "reward_std": 0.4744189381599426, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6479687690734863, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 2107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 279.125, "completions/mean_terminated_length": 279.125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.4216, "grad_norm": 0.2309381067752838, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24586956.0, "reward": 1.2512288093566895, "reward_std": 0.44498634338378906, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7524546384811401, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.995312511920929, "step": 2108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 723.0625, "completions/mean_terminated_length": 703.0000610351562, "completions/min_length": 536.0, "completions/min_terminated_length": 536.0, "epoch": 0.4218, "grad_norm": 0.09697584062814713, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24602317.0, "reward": 1.1064603328704834, "reward_std": 0.5386605262756348, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.650418758392334, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 2109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 312.5625, "completions/mean_terminated_length": 312.5625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.422, "grad_norm": 0.2539781630039215, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24611366.0, "reward": 1.0076074600219727, "reward_std": 0.508010983467102, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5152124762535095, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 379.0, "completions/max_terminated_length": 379.0, "completions/mean_length": 230.25, "completions/mean_terminated_length": 230.25, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.4222, "grad_norm": 0.22815711796283722, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24618906.0, "reward": 1.0842853784561157, "reward_std": 0.4857366383075714, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6060687303543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 539.75, "completions/mean_terminated_length": 539.75, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.4224, "grad_norm": 0.05182967334985733, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24631478.0, "reward": 0.6214760541915894, "reward_std": 0.18157246708869934, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24294999241828918, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8199999928474426, "step": 2112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1003.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 743.6875, "completions/mean_terminated_length": 743.6875, "completions/min_length": 534.0, "completions/min_terminated_length": 534.0, "epoch": 0.4226, "grad_norm": 0.11646868288516998, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24648369.0, "reward": 0.8481259346008301, "reward_std": 0.4529927372932434, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3837500214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 601.5, "completions/mean_terminated_length": 601.5, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.4228, "grad_norm": 0.08430741727352142, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24662441.0, "reward": 1.134110450744629, "reward_std": 0.4876353442668915, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6432187557220459, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 2114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 839.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 533.625, "completions/mean_terminated_length": 533.625, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.423, "grad_norm": 0.08639384061098099, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24674563.0, "reward": 1.093894362449646, "reward_std": 0.35403192043304443, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.8127874732017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6237499713897705, "step": 2115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 253.3125, "completions/mean_terminated_length": 253.3125, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.4232, "grad_norm": 0.2426312416791916, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24682120.0, "reward": 0.7755728363990784, "reward_std": 0.432017058134079, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30114373564720154, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 2116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 452.875, "completions/mean_terminated_length": 452.875, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.4234, "grad_norm": 0.1672537475824356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24693142.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 827.8125, "completions/mean_terminated_length": 710.1000366210938, "completions/min_length": 501.0, "completions/min_terminated_length": 501.0, "epoch": 0.4236, "grad_norm": 0.0733180120587349, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24710291.0, "reward": 0.4149194061756134, "reward_std": 0.4045153558254242, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14233750104904175, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5712499618530273, "step": 2118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1005.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 680.9375, "completions/mean_terminated_length": 680.9375, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.4238, "grad_norm": 0.0213934313505888, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24724914.0, "reward": 0.5820321440696716, "reward_std": 0.0935433954000473, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1640625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 2119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 501.75, "completions/mean_terminated_length": 466.933349609375, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.424, "grad_norm": 0.17126159369945526, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24737198.0, "reward": 1.0127227306365967, "reward_std": 0.5327652096748352, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5879437923431396, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8518750071525574, "step": 2120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 501.9375, "completions/mean_terminated_length": 501.9375, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.4242, "grad_norm": 0.11565564572811127, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24748877.0, "reward": 1.1370667219161987, "reward_std": 0.4836580157279968, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6491312384605408, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 2121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 871.25, "completions/mean_terminated_length": 849.4285888671875, "completions/min_length": 590.0, "completions/min_terminated_length": 590.0, "epoch": 0.4244, "grad_norm": 0.09119883179664612, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24766577.0, "reward": 0.6459386348724365, "reward_std": 0.4599021077156067, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22937500476837158, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.84375, "step": 2122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 654.1875, "completions/mean_terminated_length": 654.1875, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.4246, "grad_norm": 0.13477452099323273, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24780500.0, "reward": 1.081819772720337, "reward_std": 0.48935651779174805, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6011375188827515, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 416.125, "completions/mean_terminated_length": 416.125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.4248, "grad_norm": 0.014403731562197208, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24792310.0, "reward": 0.5441418886184692, "reward_std": 0.05192448943853378, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08828124403953552, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 688.3125, "completions/mean_terminated_length": 576.4166870117188, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.425, "grad_norm": 0.1809132695198059, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24807067.0, "reward": 0.8159040212631226, "reward_std": 0.653057873249054, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.44430625438690186, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7331249713897705, "step": 2125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 774.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 617.6875, "completions/mean_terminated_length": 617.6875, "completions/min_length": 484.0, "completions/min_terminated_length": 484.0, "epoch": 0.4252, "grad_norm": 0.14998969435691833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24820822.0, "reward": 1.3744635581970215, "reward_std": 0.3875456750392914, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9364250302314758, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9275000095367432, "step": 2126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 443.75, "completions/mean_terminated_length": 443.75, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.4254, "grad_norm": 0.1547415405511856, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24833362.0, "reward": 0.8554693460464478, "reward_std": 0.35172975063323975, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6484375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5062500238418579, "step": 2127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 837.0, "completions/mean_length": 627.8125, "completions/mean_terminated_length": 601.4000244140625, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.4256, "grad_norm": 0.07374301552772522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24847775.0, "reward": 0.8783599138259888, "reward_std": 0.3664186894893646, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6942187547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.453125, "step": 2128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 846.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 558.625, "completions/mean_terminated_length": 558.625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.4258, "grad_norm": 0.10490789264440536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24860553.0, "reward": 0.8760102987289429, "reward_std": 0.4415719509124756, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4395187497138977, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143749475479126, "step": 2129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 789.25, "completions/mean_terminated_length": 773.6000366210938, "completions/min_length": 578.0, "completions/min_terminated_length": 578.0, "epoch": 0.426, "grad_norm": 0.02316736988723278, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24877789.0, "reward": 0.5349227786064148, "reward_std": 0.15745548903942108, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13234375417232513, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 2130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 623.125, "completions/mean_terminated_length": 623.125, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.4262, "grad_norm": 0.04243868216872215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24891855.0, "reward": 0.8644536137580872, "reward_std": 0.1863705962896347, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7289062738418579, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3843750059604645, "step": 2131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 333.375, "completions/mean_terminated_length": 333.375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.4264, "grad_norm": 0.12791067361831665, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24901029.0, "reward": 1.4177324771881104, "reward_std": 0.23172813653945923, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.8979624509811401, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 2132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 436.875, "completions/mean_terminated_length": 436.875, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.4266, "grad_norm": 0.16152149438858032, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24911619.0, "reward": 1.0532820224761963, "reward_std": 0.46148309111595154, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5440624952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 2133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 314.875, "completions/mean_terminated_length": 314.875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.4268, "grad_norm": 0.17784132063388824, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24920433.0, "reward": 1.261404037475586, "reward_std": 0.4249444901943207, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7728062868118286, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 423.5, "completions/mean_terminated_length": 423.5, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.427, "grad_norm": 0.14321434497833252, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24931337.0, "reward": 1.1886605024337769, "reward_std": 0.47537216544151306, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.689818799495697, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 2135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 410.0, "completions/max_terminated_length": 410.0, "completions/mean_length": 330.5625, "completions/mean_terminated_length": 330.5625, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.4272, "grad_norm": 0.0783267542719841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24940994.0, "reward": 1.4420322179794312, "reward_std": 0.22587692737579346, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9465624690055847, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 488.9375, "completions/mean_terminated_length": 488.9375, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.4274, "grad_norm": 0.09682878851890564, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24952641.0, "reward": 1.2759979963302612, "reward_std": 0.399666428565979, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.801993727684021, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 2137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 268.0, "completions/max_terminated_length": 268.0, "completions/mean_length": 228.0, "completions/mean_terminated_length": 228.0, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.4276, "grad_norm": 0.2543591260910034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24960049.0, "reward": 1.317732334136963, "reward_std": 0.3897194564342499, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8229625225067139, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 568.1875, "completions/mean_terminated_length": 537.800048828125, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.4278, "grad_norm": 0.0930367261171341, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24972692.0, "reward": 0.8976445198059082, "reward_std": 0.4471990466117859, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6077874898910522, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6174999475479126, "step": 2139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 605.5625, "completions/mean_terminated_length": 545.7857666015625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.428, "grad_norm": 0.037987060844898224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24986285.0, "reward": 0.46554484963417053, "reward_std": 0.18302902579307556, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05608749762177467, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 2140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 770.625, "completions/mean_terminated_length": 618.6000366210938, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.4282, "grad_norm": 0.10573609918355942, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25002423.0, "reward": 0.37265685200691223, "reward_std": 0.3165701627731323, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12031249701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.543749988079071, "step": 2141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 519.0, "completions/mean_terminated_length": 519.0, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.4284, "grad_norm": 0.12300895154476166, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25014359.0, "reward": 1.2327353954315186, "reward_std": 0.41998428106307983, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7779687643051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 2142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 593.0625, "completions/mean_terminated_length": 493.61541748046875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.4286, "grad_norm": 0.08080411702394485, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25028552.0, "reward": 0.5149886608123779, "reward_std": 0.36431270837783813, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15497498214244843, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7425000071525574, "step": 2143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 317.9375, "completions/mean_terminated_length": 317.9375, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.4288, "grad_norm": 0.15344424545764923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25037895.0, "reward": 1.2002854347229004, "reward_std": 0.4577510356903076, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7130687236785889, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 2144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 746.875, "completions/mean_terminated_length": 728.4000244140625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.429, "grad_norm": 0.05284969508647919, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25053973.0, "reward": 0.7514036893844604, "reward_std": 0.3443158268928528, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5028062462806702, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5556249618530273, "step": 2145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 555.875, "completions/mean_terminated_length": 555.875, "completions/min_length": 456.0, "completions/min_terminated_length": 456.0, "epoch": 0.4292, "grad_norm": 0.04120872542262077, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25066499.0, "reward": 1.4394540786743164, "reward_std": 0.23752106726169586, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.94140625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 960.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 674.0625, "completions/mean_terminated_length": 674.0625, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.4294, "grad_norm": 0.11167514324188232, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25081012.0, "reward": 1.0293633937835693, "reward_std": 0.48017293214797974, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5587249994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 2147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 468.25, "completions/mean_terminated_length": 468.25, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.4296, "grad_norm": 0.1225445345044136, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25092408.0, "reward": 0.7785290479660034, "reward_std": 0.4299801290035248, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3070562481880188, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 870.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 547.375, "completions/mean_terminated_length": 547.375, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.4298, "grad_norm": 0.06756394356489182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25105598.0, "reward": 1.336094617843628, "reward_std": 0.3028666079044342, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.9221875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7999999523162842, "step": 2149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 314.5625, "completions/mean_terminated_length": 314.5625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.43, "grad_norm": 0.13175645470619202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25115383.0, "reward": 1.257763385772705, "reward_std": 0.43146947026252747, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7655249834060669, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 795.875, "completions/mean_terminated_length": 763.2857666015625, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.4302, "grad_norm": 0.0911736860871315, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25132085.0, "reward": 0.7907695174217224, "reward_std": 0.4474034607410431, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39403748512268066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7987499833106995, "step": 2151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 629.1875, "completions/mean_terminated_length": 629.1875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.4304, "grad_norm": 0.056546151638031006, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25146936.0, "reward": 0.8264849185943604, "reward_std": 0.2167346179485321, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6529687643051147, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.421875, "step": 2152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 520.375, "completions/mean_terminated_length": 520.375, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.4306, "grad_norm": 0.14803609251976013, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25159342.0, "reward": 1.2592477798461914, "reward_std": 0.42953038215637207, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7684937715530396, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 2153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 424.6875, "completions/mean_terminated_length": 384.7333679199219, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.4308, "grad_norm": 0.12450668960809708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25169801.0, "reward": 1.1935945749282837, "reward_std": 0.49964940547943115, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7621874809265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "step": 2154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 348.6875, "completions/mean_terminated_length": 348.6875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.431, "grad_norm": 0.17522761225700378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25179668.0, "reward": 1.1392853260040283, "reward_std": 0.4432883560657501, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7160687446594238, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 2155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 366.0, "completions/max_terminated_length": 366.0, "completions/mean_length": 297.4375, "completions/mean_terminated_length": 297.4375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.4312, "grad_norm": 0.05713041499257088, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25187899.0, "reward": 1.4399102926254272, "reward_std": 0.23764310777187347, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9423187971115112, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 2156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 232.0, "completions/max_terminated_length": 232.0, "completions/mean_length": 156.6875, "completions/mean_terminated_length": 156.6875, "completions/min_length": 109.0, "completions/min_terminated_length": 109.0, "epoch": 0.4314, "grad_norm": 1.0934925740002654e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25193894.0, "reward": 1.4999891519546509, "reward_std": 2.278201463923324e-05, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9999749660491943, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9975000023841858, "step": 2157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 489.875, "completions/mean_terminated_length": 454.2666931152344, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.4316, "grad_norm": 0.14647126197814941, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25206148.0, "reward": 1.0411696434020996, "reward_std": 0.5370585918426514, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.582337498664856, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8650000095367432, "step": 2158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 702.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 478.5625, "completions/mean_terminated_length": 478.5625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.4318, "grad_norm": 0.11216998845338821, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25217997.0, "reward": 0.8121883869171143, "reward_std": 0.3929617404937744, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37437498569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 2159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 425.375, "completions/mean_terminated_length": 425.375, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.432, "grad_norm": 0.23072603344917297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25228755.0, "reward": 0.9559979438781738, "reward_std": 0.49493661522865295, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47449374198913574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 702.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 522.0625, "completions/mean_terminated_length": 522.0625, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.4322, "grad_norm": 0.11082924902439117, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25242084.0, "reward": 0.9548321962356567, "reward_std": 0.3819458484649658, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6596624851226807, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5987499952316284, "step": 2161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 318.5625, "completions/mean_terminated_length": 318.5625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.4324, "grad_norm": 0.21085308492183685, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25251277.0, "reward": 1.023982286453247, "reward_std": 0.49081704020500183, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5479625463485718, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587499499320984, "step": 2162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 524.5625, "completions/mean_terminated_length": 524.5625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.4326, "grad_norm": 0.08567292988300323, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25263462.0, "reward": 0.6552227735519409, "reward_std": 0.19625715911388397, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3104437589645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7518749833106995, "step": 2163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 935.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 692.75, "completions/mean_terminated_length": 692.75, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.4328, "grad_norm": 0.03993246704339981, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25279442.0, "reward": 0.8989067077636719, "reward_std": 0.17195113003253937, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7978124618530273, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.30000001192092896, "step": 2164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 319.1875, "completions/mean_terminated_length": 319.1875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.433, "grad_norm": 0.1392437070608139, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25288325.0, "reward": 1.4401042461395264, "reward_std": 0.2376948595046997, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9427062273025513, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 2165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 374.0, "completions/mean_terminated_length": 374.0, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.4332, "grad_norm": 0.005408462602645159, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25298965.0, "reward": 0.5438947677612305, "reward_std": 0.013753141276538372, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08778749406337738, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 2166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 575.875, "completions/mean_terminated_length": 575.875, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.4334, "grad_norm": 0.010223576799035072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25312131.0, "reward": 0.5593758821487427, "reward_std": 0.04249993711709976, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11874999105930328, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 2167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 291.8125, "completions/mean_terminated_length": 291.8125, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.4336, "grad_norm": 0.08167114853858948, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25320480.0, "reward": 1.437472939491272, "reward_std": 0.24733991920948029, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.937443733215332, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 2168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 488.875, "completions/mean_terminated_length": 488.875, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.4338, "grad_norm": 0.05420216545462608, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25332894.0, "reward": 0.5905352830886841, "reward_std": 0.13117870688438416, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18106874823570251, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8831249475479126, "step": 2169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 836.8125, "completions/mean_terminated_length": 824.3333740234375, "completions/min_length": 662.0, "completions/min_terminated_length": 662.0, "epoch": 0.434, "grad_norm": 0.07059436291456223, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25351003.0, "reward": 0.9920194149017334, "reward_std": 0.398874431848526, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.796537458896637, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.44875001907348633, "step": 2170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 460.5, "completions/mean_terminated_length": 460.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.4342, "grad_norm": 0.0043232799507677555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25362803.0, "reward": 0.5402323007583618, "reward_std": 0.018434172496199608, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0804625004529953, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 510.625, "completions/mean_terminated_length": 510.625, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.4344, "grad_norm": 0.08646956831216812, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25375933.0, "reward": 0.6685822010040283, "reward_std": 0.32436856627464294, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21216249465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 417.0, "completions/mean_terminated_length": 417.0, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.4346, "grad_norm": 0.1416904479265213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25386157.0, "reward": 1.3786041736602783, "reward_std": 0.32963475584983826, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8822062611579895, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 299.125, "completions/mean_terminated_length": 299.125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.4348, "grad_norm": 0.2217225581407547, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25395263.0, "reward": 1.2031760215759277, "reward_std": 0.45348861813545227, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7188500165939331, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 566.1875, "completions/mean_terminated_length": 566.1875, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.435, "grad_norm": 0.14584404230117798, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25408498.0, "reward": 1.1332038640975952, "reward_std": 0.43834900856018066, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.8289061784744263, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.621874988079071, "step": 2175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 545.375, "completions/mean_terminated_length": 545.375, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.4352, "grad_norm": 0.07961202412843704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25421880.0, "reward": 0.8076447248458862, "reward_std": 0.425613135099411, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.36528751254081726, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 2176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 677.1875, "completions/mean_terminated_length": 677.1875, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.4354, "grad_norm": 0.12380277365446091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25437339.0, "reward": 1.2174384593963623, "reward_std": 0.436681866645813, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7473750114440918, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 764.625, "completions/mean_terminated_length": 747.3333740234375, "completions/min_length": 507.0, "completions/min_terminated_length": 507.0, "epoch": 0.4356, "grad_norm": 0.06610694527626038, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25454213.0, "reward": 0.8393757343292236, "reward_std": 0.41980600357055664, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5537500381469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.625, "step": 2178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 475.0, "completions/max_terminated_length": 475.0, "completions/mean_length": 339.875, "completions/mean_terminated_length": 339.875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.4358, "grad_norm": 0.21290495991706848, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25463779.0, "reward": 1.1691135168075562, "reward_std": 0.45302873849868774, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7132250070571899, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9000000357627869, "step": 2179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 429.8125, "completions/mean_terminated_length": 429.8125, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.436, "grad_norm": 0.1342644989490509, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25474640.0, "reward": 1.2000665664672852, "reward_std": 0.457598477602005, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7126312255859375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 596.0, "completions/mean_terminated_length": 534.857177734375, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.4362, "grad_norm": 0.1243230402469635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25487792.0, "reward": 0.9122661352157593, "reward_std": 0.45619261264801025, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6995312571525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.47187501192092896, "step": 2181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 371.0625, "completions/mean_terminated_length": 371.0625, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.4364, "grad_norm": 0.09640492498874664, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25497697.0, "reward": 1.4395259618759155, "reward_std": 0.2375403493642807, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415500164031982, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 438.75, "completions/mean_terminated_length": 438.75, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.4366, "grad_norm": 0.006541526410728693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25509165.0, "reward": 0.5408541560173035, "reward_std": 0.0171066727489233, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08170624822378159, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 395.4375, "completions/mean_terminated_length": 395.4375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.4368, "grad_norm": 0.10038898885250092, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25519284.0, "reward": 0.5995010137557983, "reward_std": 0.2409001737833023, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13650000095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 284.1875, "completions/mean_terminated_length": 284.1875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.437, "grad_norm": 0.0837346762418747, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25527575.0, "reward": 1.3805947303771973, "reward_std": 0.324719101190567, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8861874938011169, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 732.0, "completions/mean_terminated_length": 690.2857666015625, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.4372, "grad_norm": 0.11220482736825943, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25543751.0, "reward": 0.9517196416854858, "reward_std": 0.5585231781005859, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5909374952316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.71875, "step": 2186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 603.6875, "completions/mean_terminated_length": 603.6875, "completions/min_length": 436.0, "completions/min_terminated_length": 436.0, "epoch": 0.4374, "grad_norm": 0.004955068230628967, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25559042.0, "reward": 0.5244264006614685, "reward_std": 0.02305559813976288, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04884999990463257, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 923.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 575.5625, "completions/mean_terminated_length": 575.5625, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.4376, "grad_norm": 0.05993428826332092, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25572363.0, "reward": 1.3643509149551392, "reward_std": 0.3025941252708435, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9162000417709351, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 2188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 521.8125, "completions/mean_terminated_length": 488.3333740234375, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.4378, "grad_norm": 0.11545717716217041, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25585336.0, "reward": 0.6807533502578735, "reward_std": 0.42718505859375, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2365046739578247, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9028124809265137, "step": 2189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 902.25, "completions/mean_terminated_length": 846.9091186523438, "completions/min_length": 734.0, "completions/min_terminated_length": 734.0, "epoch": 0.438, "grad_norm": 0.11106304079294205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25604620.0, "reward": 0.6666007041931152, "reward_std": 0.6224759817123413, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33319997787475586, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6725000143051147, "step": 2190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 721.5625, "completions/mean_terminated_length": 701.4000244140625, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.4382, "grad_norm": 0.12927985191345215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25620437.0, "reward": 0.8460947275161743, "reward_std": 0.5569004416465759, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44218751788139343, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 2191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 642.375, "completions/mean_terminated_length": 616.933349609375, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.4384, "grad_norm": 0.017695769667625427, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25635467.0, "reward": 0.51734459400177, "reward_std": 0.1384401023387909, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 2192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 353.1875, "completions/mean_terminated_length": 353.1875, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.4386, "grad_norm": 0.00021971725800540298, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25644942.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 434.6875, "completions/mean_terminated_length": 395.4000244140625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.4388, "grad_norm": 0.10176318138837814, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25655593.0, "reward": 0.6816291809082031, "reward_std": 0.4269339442253113, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23825624585151672, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 2194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 527.9375, "completions/mean_terminated_length": 527.9375, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.439, "grad_norm": 0.1459413468837738, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25668056.0, "reward": 0.9646104574203491, "reward_std": 0.48700499534606934, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49171876907348633, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 2195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 840.0625, "completions/mean_terminated_length": 778.75, "completions/min_length": 523.0, "completions/min_terminated_length": 523.0, "epoch": 0.4392, "grad_norm": 0.09194622933864594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25685465.0, "reward": 0.5518753528594971, "reward_std": 0.41948986053466797, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.41624999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.375, "step": 2196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 486.875, "completions/mean_terminated_length": 486.875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.4394, "grad_norm": 0.13014332950115204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25697735.0, "reward": 0.8227353096008301, "reward_std": 0.4166373014450073, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39546874165534973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 2197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 404.25, "completions/mean_terminated_length": 404.25, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.4396, "grad_norm": 0.14331813156604767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25709035.0, "reward": 1.076079249382019, "reward_std": 0.49565792083740234, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5896562337875366, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 2198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 761.4375, "completions/mean_terminated_length": 723.9285888671875, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.4398, "grad_norm": 0.06951609253883362, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25725442.0, "reward": 0.5421884059906006, "reward_std": 0.24899958074092865, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20937499403953552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 2199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 243.4375, "completions/mean_terminated_length": 243.4375, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.44, "grad_norm": 0.2820882797241211, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25733017.0, "reward": 1.286163568496704, "reward_std": 0.3509472608566284, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8848249912261963, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8012499809265137, "step": 2200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 869.0, "completions/max_terminated_length": 869.0, "completions/mean_length": 524.4375, "completions/mean_terminated_length": 524.4375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.4402, "grad_norm": 0.05476420745253563, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25745312.0, "reward": 0.5917043685913086, "reward_std": 0.24287573993206024, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12090624868869781, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 813.5625, "completions/mean_terminated_length": 765.0, "completions/min_length": 555.0, "completions/min_terminated_length": 555.0, "epoch": 0.4404, "grad_norm": 0.08317404240369797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25762841.0, "reward": 0.7272505760192871, "reward_std": 0.49739378690719604, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4544999897480011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6137499809265137, "step": 2202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 863.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 594.25, "completions/mean_terminated_length": 594.25, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.4406, "grad_norm": 0.08206293731927872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25776685.0, "reward": 0.7832039594650269, "reward_std": 0.2767755091190338, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.50390625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 2203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 537.8125, "completions/mean_terminated_length": 537.8125, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.4408, "grad_norm": 0.18694274127483368, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25789610.0, "reward": 1.0220727920532227, "reward_std": 0.4926318824291229, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.544143795967102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 435.1875, "completions/mean_terminated_length": 435.1875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.441, "grad_norm": 0.13067291676998138, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25801005.0, "reward": 1.2818571329116821, "reward_std": 0.3952958583831787, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.813712477684021, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287500381469727, "step": 2205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 849.25, "completions/mean_terminated_length": 769.8181762695312, "completions/min_length": 568.0, "completions/min_terminated_length": 568.0, "epoch": 0.4412, "grad_norm": 0.09525793045759201, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25818513.0, "reward": 0.6899224519729614, "reward_std": 0.6188130974769592, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3798437714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 2206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 353.1875, "completions/mean_terminated_length": 353.1875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.4414, "grad_norm": 0.2160319983959198, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25828276.0, "reward": 0.8454697132110596, "reward_std": 0.4552173614501953, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37843751907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 647.1875, "completions/mean_terminated_length": 622.0667114257812, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.4416, "grad_norm": 0.03485242277383804, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25842855.0, "reward": 0.5257540345191956, "reward_std": 0.16424833238124847, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11400624364614487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.871874988079071, "step": 2208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 520.6875, "completions/mean_terminated_length": 520.6875, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.4418, "grad_norm": 0.07826340198516846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25854722.0, "reward": 0.6931914687156677, "reward_std": 0.2692176103591919, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.32388123869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8081250190734863, "step": 2209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 387.375, "completions/mean_terminated_length": 387.375, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.442, "grad_norm": 0.2524963915348053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25865368.0, "reward": 1.0237510204315186, "reward_std": 0.49057790637016296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 933.0, "completions/mean_length": 636.6875, "completions/mean_terminated_length": 610.86669921875, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.4422, "grad_norm": 0.027590680867433548, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25879571.0, "reward": 0.5202353000640869, "reward_std": 0.1396174281835556, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10296875238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 2211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 409.9375, "completions/mean_terminated_length": 409.9375, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.4424, "grad_norm": 0.1528734564781189, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25890498.0, "reward": 1.1289575099945068, "reward_std": 0.4941324293613434, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6329125165939331, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 2212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 385.25, "completions/mean_terminated_length": 385.25, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.4426, "grad_norm": 0.1582842767238617, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25900310.0, "reward": 0.982735276222229, "reward_std": 0.4258478581905365, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5279687643051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8093750476837158, "step": 2213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 303.9375, "completions/mean_terminated_length": 303.9375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.4428, "grad_norm": 0.15129545331001282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25908949.0, "reward": 1.3145699501037598, "reward_std": 0.3975796103477478, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8166375160217285, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 2214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 335.5625, "completions/mean_terminated_length": 289.66668701171875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.443, "grad_norm": 0.3334440588951111, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25917982.0, "reward": 1.1855355501174927, "reward_std": 0.5078167915344238, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7460687160491943, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8881250619888306, "step": 2215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 426.4375, "completions/mean_terminated_length": 386.6000061035156, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.4432, "grad_norm": 0.14876170456409454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25929221.0, "reward": 1.2552165985107422, "reward_std": 0.47301042079925537, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.822931170463562, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8506250381469727, "step": 2216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 538.6875, "completions/mean_terminated_length": 538.6875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.4434, "grad_norm": 0.12922082841396332, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25942032.0, "reward": 1.0722664594650269, "reward_std": 0.4411281645298004, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.64453125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.840624988079071, "step": 2217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 332.4375, "completions/mean_terminated_length": 332.4375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.4436, "grad_norm": 0.1737278252840042, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25950759.0, "reward": 0.8432759046554565, "reward_std": 0.4566813111305237, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37404999136924744, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 604.9375, "completions/mean_terminated_length": 604.9375, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.4438, "grad_norm": 0.07740175724029541, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25964342.0, "reward": 0.6067821979522705, "reward_std": 0.23848137259483337, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1510625034570694, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 2219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 572.0, "completions/mean_terminated_length": 467.69232177734375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.444, "grad_norm": 0.0862201675772667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25977622.0, "reward": 0.6526572704315186, "reward_std": 0.4439195394515991, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.36781251430511475, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5125000476837158, "step": 2220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 856.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 551.1875, "completions/mean_terminated_length": 551.1875, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.4442, "grad_norm": 0.09786121547222137, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25990105.0, "reward": 0.8123445510864258, "reward_std": 0.3580760359764099, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4371874928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7874999642372131, "step": 2221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 406.9375, "completions/mean_terminated_length": 406.9375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.4444, "grad_norm": 0.13808277249336243, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26000440.0, "reward": 1.3209104537963867, "reward_std": 0.38309869170188904, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8293187618255615, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 2222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 339.5625, "completions/mean_terminated_length": 339.5625, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.4446, "grad_norm": 0.1958799511194229, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26009745.0, "reward": 1.256525993347168, "reward_std": 0.43374940752983093, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7630499601364136, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 488.625, "completions/mean_terminated_length": 488.625, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.4448, "grad_norm": 0.05987563729286194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26022139.0, "reward": 0.602719783782959, "reward_std": 0.2396908849477768, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14293749630451202, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 516.1875, "completions/mean_terminated_length": 516.1875, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.445, "grad_norm": 0.1570957899093628, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26034302.0, "reward": 1.2649807929992676, "reward_std": 0.40523597598075867, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7799596786499023, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9129166603088379, "step": 2225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 576.5625, "completions/mean_terminated_length": 576.5625, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.4452, "grad_norm": 0.09524957835674286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26047543.0, "reward": 1.261447787284851, "reward_std": 0.42497050762176514, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7728937268257141, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 605.6875, "completions/mean_terminated_length": 605.6875, "completions/min_length": 500.0, "completions/min_terminated_length": 500.0, "epoch": 0.4454, "grad_norm": 0.1010618507862091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26061202.0, "reward": 0.8001572489738464, "reward_std": 0.4222693145275116, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3503125011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 2227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 447.0, "completions/mean_terminated_length": 447.0, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.4456, "grad_norm": 0.10421404242515564, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26072306.0, "reward": 1.2595915794372559, "reward_std": 0.42392992973327637, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691812515258789, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9443750381469727, "step": 2228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 398.125, "completions/mean_terminated_length": 398.125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.4458, "grad_norm": 0.21123291552066803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26082100.0, "reward": 1.0238291025161743, "reward_std": 0.4906587302684784, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.547656238079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 321.0625, "completions/mean_terminated_length": 321.0625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.446, "grad_norm": 0.19633722305297852, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26090757.0, "reward": 1.138169765472412, "reward_std": 0.4820207357406616, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6513375043869019, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 2230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 344.9375, "completions/mean_terminated_length": 344.9375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4462, "grad_norm": 0.10211789608001709, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26101124.0, "reward": 1.3911478519439697, "reward_std": 0.2973497807979584, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9072937965393066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9381250143051147, "step": 2231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 376.6875, "completions/mean_terminated_length": 376.6875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.4464, "grad_norm": 0.07606273889541626, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26110799.0, "reward": 1.364532232284546, "reward_std": 0.32118460536003113, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8540624976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 2232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 684.6875, "completions/mean_terminated_length": 662.0667114257812, "completions/min_length": 534.0, "completions/min_terminated_length": 534.0, "epoch": 0.4466, "grad_norm": 0.06436891853809357, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26125562.0, "reward": 0.7625004649162292, "reward_std": 0.2662461996078491, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5875000357627869, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4937500059604645, "step": 2233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 262.0, "completions/max_terminated_length": 262.0, "completions/mean_length": 213.4375, "completions/mean_terminated_length": 213.4375, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.4468, "grad_norm": 0.12705230712890625, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26132689.0, "reward": 1.4374480247497559, "reward_std": 0.24998557567596436, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9373937845230103, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9906250238418579, "step": 2234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 398.125, "completions/mean_terminated_length": 398.125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.447, "grad_norm": 0.15315426886081696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26142963.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 822.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 619.25, "completions/mean_terminated_length": 619.25, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.4472, "grad_norm": 0.0029852723237127066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26157367.0, "reward": 0.551641583442688, "reward_std": 0.011562498286366463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10328124463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 280.9375, "completions/mean_terminated_length": 280.9375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.4474, "grad_norm": 0.26177921891212463, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26165526.0, "reward": 1.1380259990692139, "reward_std": 0.4819057881832123, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6510499715805054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 709.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 444.8125, "completions/mean_terminated_length": 444.8125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.4476, "grad_norm": 0.14498867094516754, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26176291.0, "reward": 0.9458823204040527, "reward_std": 0.5044989585876465, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45426249504089355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 2238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 794.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 516.125, "completions/mean_terminated_length": 516.125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.4478, "grad_norm": 0.06099151447415352, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26190037.0, "reward": 0.6092416048049927, "reward_std": 0.17457731068134308, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2184811234474182, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8317500352859497, "step": 2239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 546.6875, "completions/mean_terminated_length": 546.6875, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.448, "grad_norm": 0.14512866735458374, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26202912.0, "reward": 0.9008418321609497, "reward_std": 0.44440290331840515, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.489181250333786, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8568750619888306, "step": 2240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 742.375, "completions/mean_terminated_length": 702.1428833007812, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.4482, "grad_norm": 0.08650421351194382, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26220262.0, "reward": 0.8346096277236938, "reward_std": 0.3479450047016144, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7942187786102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.171875, "step": 2241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 704.9375, "completions/mean_terminated_length": 631.3077392578125, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.4484, "grad_norm": 0.1237831860780716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26237013.0, "reward": 0.8590601682662964, "reward_std": 0.6154521107673645, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46811872720718384, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7743749618530273, "step": 2242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 354.0, "completions/max_terminated_length": 354.0, "completions/mean_length": 239.1875, "completions/mean_terminated_length": 239.1875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.4486, "grad_norm": 0.00028674848726950586, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26244904.0, "reward": 1.4990227222442627, "reward_std": 0.00048751343274489045, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980437755584717, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 916.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 451.25, "completions/mean_terminated_length": 451.25, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.4488, "grad_norm": 0.14704042673110962, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26255980.0, "reward": 0.8446882963180542, "reward_std": 0.40280142426490784, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5643749833106995, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5625, "step": 2244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 932.25, "completions/mean_terminated_length": 840.5, "completions/min_length": 677.0, "completions/min_terminated_length": 677.0, "epoch": 0.449, "grad_norm": 0.09631413221359253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26274976.0, "reward": 0.6583472490310669, "reward_std": 0.7223047018051147, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4416937530040741, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4206250011920929, "step": 2245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 498.75, "completions/mean_terminated_length": 498.75, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.4492, "grad_norm": 0.1649560183286667, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26286652.0, "reward": 1.1220324039459229, "reward_std": 0.48036766052246094, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6190624833106995, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 2246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 272.8125, "completions/mean_terminated_length": 272.8125, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.4494, "grad_norm": 0.00023468327708542347, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26295177.0, "reward": 1.4990415573120117, "reward_std": 0.0005210168310441077, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980812668800354, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 2247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 278.0, "completions/max_terminated_length": 278.0, "completions/mean_length": 226.9375, "completions/mean_terminated_length": 226.9375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.4496, "grad_norm": 0.0005824321997351944, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26302376.0, "reward": 1.4990572929382324, "reward_std": 0.0005482406122609973, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981124997138977, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 765.5625, "completions/mean_terminated_length": 679.4166870117188, "completions/min_length": 423.0, "completions/min_terminated_length": 423.0, "epoch": 0.4498, "grad_norm": 0.13930822908878326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26318353.0, "reward": 0.6429694890975952, "reward_std": 0.5477436184883118, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28593748807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6749999523162842, "step": 2249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 444.0625, "completions/mean_terminated_length": 444.0625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.45, "grad_norm": 0.0001317009882768616, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26329090.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 484.6875, "completions/mean_terminated_length": 484.6875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.4502, "grad_norm": 0.07362274080514908, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26341645.0, "reward": 0.8415631651878357, "reward_std": 0.3202991187572479, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5581249594688416, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.643750011920929, "step": 2251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 274.6875, "completions/mean_terminated_length": 274.6875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.4504, "grad_norm": 0.19816431403160095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26349656.0, "reward": 1.2152822017669678, "reward_std": 0.43798795342445374, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7430624961853027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 2252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 392.0625, "completions/mean_terminated_length": 392.0625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4506, "grad_norm": 0.1622549146413803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26359961.0, "reward": 0.9607571363449097, "reward_std": 0.4901149272918701, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48401251435279846, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 207.0625, "completions/mean_terminated_length": 207.0625, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.4508, "grad_norm": 0.14971928298473358, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26367034.0, "reward": 0.5077884793281555, "reward_std": 0.13621856272220612, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07807499170303345, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 2254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 375.5, "completions/mean_terminated_length": 375.5, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.451, "grad_norm": 0.2146868258714676, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26377762.0, "reward": 0.9051321744918823, "reward_std": 0.47517508268356323, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4352625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 751.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 433.625, "completions/mean_terminated_length": 433.625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.4512, "grad_norm": 0.16407325863838196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26389020.0, "reward": 1.137644648551941, "reward_std": 0.39131006598472595, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7752875089645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7849999666213989, "step": 2256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 500.4375, "completions/mean_terminated_length": 500.4375, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.4514, "grad_norm": 0.0841267853975296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26400691.0, "reward": 1.4395259618759155, "reward_std": 0.2375403493642807, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415500164031982, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 531.875, "completions/mean_terminated_length": 531.875, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.4516, "grad_norm": 0.025301994755864143, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26413265.0, "reward": 0.8386724591255188, "reward_std": 0.09697005152702332, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6773437261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.546875, "step": 2258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 474.4375, "completions/mean_terminated_length": 474.4375, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.4518, "grad_norm": 0.2101612240076065, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26424344.0, "reward": 1.022085428237915, "reward_std": 0.49264490604400635, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5441687107086182, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 2259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 431.3125, "completions/mean_terminated_length": 431.3125, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.452, "grad_norm": 0.12914806604385376, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26435837.0, "reward": 0.6081259250640869, "reward_std": 0.23749999701976776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1537500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 440.875, "completions/mean_terminated_length": 440.875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.4522, "grad_norm": 0.08727473020553589, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26447163.0, "reward": 0.7106975317001343, "reward_std": 0.20531129837036133, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4213937520980835, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6768749952316284, "step": 2261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 650.75, "completions/mean_terminated_length": 650.75, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.4524, "grad_norm": 0.030822575092315674, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26462871.0, "reward": 0.6010947227478027, "reward_std": 0.11838370561599731, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20218750834465027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 2262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 513.0625, "completions/mean_terminated_length": 479.0000305175781, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.4526, "grad_norm": 0.13226568698883057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26474824.0, "reward": 1.1357040405273438, "reward_std": 0.5164752006530762, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7089062929153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 2263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 713.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 456.4375, "completions/mean_terminated_length": 456.4375, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.4528, "grad_norm": 0.09865725785493851, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26486559.0, "reward": 0.7942193746566772, "reward_std": 0.2608182430267334, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6509374976158142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4728749990463257, "step": 2264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 503.6875, "completions/mean_terminated_length": 503.6875, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.453, "grad_norm": 0.10839861631393433, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26498506.0, "reward": 1.3772916793823242, "reward_std": 0.3330632746219635, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8795812726020813, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 934.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 595.25, "completions/mean_terminated_length": 595.25, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.4532, "grad_norm": 0.13696779310703278, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26512350.0, "reward": 1.1740384101867676, "reward_std": 0.4418366551399231, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7230750322341919, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9099999666213989, "step": 2266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 526.5, "completions/mean_terminated_length": 526.5, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.4534, "grad_norm": 0.12788070738315582, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26524422.0, "reward": 0.8380824327468872, "reward_std": 0.46090731024742126, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3636624813079834, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 668.8125, "completions/mean_terminated_length": 586.84619140625, "completions/min_length": 423.0, "completions/min_terminated_length": 423.0, "epoch": 0.4536, "grad_norm": 0.16499312222003937, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26538915.0, "reward": 0.7703132629394531, "reward_std": 0.5357047319412231, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4781250059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 2268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 228.0, "completions/mean_terminated_length": 228.0, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.4538, "grad_norm": 0.30650028586387634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26546499.0, "reward": 1.3187072277069092, "reward_std": 0.38850846886634827, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8249125480651855, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 227.5625, "completions/mean_terminated_length": 227.5625, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.454, "grad_norm": 0.20064932107925415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26554188.0, "reward": 0.8668760061264038, "reward_std": 0.4462396502494812, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.42125001549720764, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9225000143051147, "step": 2270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 818.625, "completions/mean_terminated_length": 725.2727661132812, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.4542, "grad_norm": 0.19223015010356903, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26571686.0, "reward": 0.6392191052436829, "reward_std": 0.4889986515045166, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5284374952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.3125, "step": 2271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 449.25, "completions/mean_terminated_length": 449.25, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.4544, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26582586.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 988.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 555.9375, "completions/mean_terminated_length": 555.9375, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.4546, "grad_norm": 0.15283483266830444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26596153.0, "reward": 1.209607481956482, "reward_std": 0.4460946023464203, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7317124605178833, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 441.625, "completions/mean_terminated_length": 441.625, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.4548, "grad_norm": 0.16799069941043854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26607379.0, "reward": 0.6676540970802307, "reward_std": 0.3249053657054901, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21030624210834503, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 410.5, "completions/mean_terminated_length": 410.5, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.455, "grad_norm": 0.13312949240207672, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26617979.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 323.6875, "completions/mean_terminated_length": 323.6875, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.4552, "grad_norm": 0.25258681178092957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26626454.0, "reward": 1.197997808456421, "reward_std": 0.4616157114505768, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.708493709564209, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 2276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 412.1875, "completions/mean_terminated_length": 412.1875, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.4554, "grad_norm": 0.17112278938293457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26636873.0, "reward": 1.0262444019317627, "reward_std": 0.372693806886673, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7399874925613403, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6299999952316284, "step": 2277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 476.25, "completions/mean_terminated_length": 476.25, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.4556, "grad_norm": 0.249245747923851, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26648525.0, "reward": 0.7042196393013, "reward_std": 0.3298622965812683, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28343749046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 2278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 942.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 516.625, "completions/mean_terminated_length": 516.625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.4558, "grad_norm": 0.14592106640338898, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26660279.0, "reward": 1.0219416618347168, "reward_std": 0.4924962520599365, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5438812971115112, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 808.6875, "completions/mean_terminated_length": 759.0, "completions/min_length": 511.0, "completions/min_terminated_length": 511.0, "epoch": 0.456, "grad_norm": 0.1379602700471878, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26677282.0, "reward": 0.7048443555831909, "reward_std": 0.4958288073539734, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.40968751907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.643750011920929, "step": 2280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 217.0, "completions/max_terminated_length": 217.0, "completions/mean_length": 162.0, "completions/mean_terminated_length": 162.0, "completions/min_length": 124.0, "completions/min_terminated_length": 124.0, "epoch": 0.4562, "grad_norm": 0.1738528311252594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26683762.0, "reward": 1.4398291110992432, "reward_std": 0.23762162029743195, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421563148498535, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 2281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 600.5, "completions/mean_terminated_length": 540.0, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.4564, "grad_norm": 0.13044771552085876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26697306.0, "reward": 0.574063241481781, "reward_std": 0.2831222712993622, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2731249928474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.668749988079071, "step": 2282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 366.0, "completions/mean_terminated_length": 366.0, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.4566, "grad_norm": 0.06868495047092438, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26707370.0, "reward": 1.4402449131011963, "reward_std": 0.2377321571111679, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9429875016212463, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 2283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 329.5625, "completions/mean_terminated_length": 329.5625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.4568, "grad_norm": 0.08025658130645752, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26716195.0, "reward": 0.8217978477478027, "reward_std": 0.21297062933444977, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6435937881469727, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.421875, "step": 2284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 583.25, "completions/mean_terminated_length": 583.25, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.457, "grad_norm": 0.12745703756809235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26729287.0, "reward": 0.92484450340271, "reward_std": 0.34567883610725403, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6621874570846558, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.574999988079071, "step": 2285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 495.0, "completions/mean_terminated_length": 495.0, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.4572, "grad_norm": 0.03837599977850914, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26741447.0, "reward": 0.5905323028564453, "reward_std": 0.10857240855693817, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1810624897480011, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8949999809265137, "step": 2286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 586.0, "completions/mean_terminated_length": 586.0, "completions/min_length": 454.0, "completions/min_terminated_length": 454.0, "epoch": 0.4574, "grad_norm": 0.05546649545431137, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26758359.0, "reward": 0.9523441791534424, "reward_std": 0.192862406373024, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.8421875238418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 2287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 696.5, "completions/mean_terminated_length": 587.3333740234375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.4576, "grad_norm": 0.09580866247415543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26773823.0, "reward": 0.7010939717292786, "reward_std": 0.4347938299179077, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6521874666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.17499999701976776, "step": 2288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 613.6875, "completions/mean_terminated_length": 555.0714721679688, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.4578, "grad_norm": 0.1465553641319275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26787738.0, "reward": 0.8226568698883057, "reward_std": 0.42445048689842224, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.645312488079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.41874998807907104, "step": 2289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 791.0, "completions/max_terminated_length": 791.0, "completions/mean_length": 340.5, "completions/mean_terminated_length": 340.5, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.458, "grad_norm": 0.24514955282211304, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26796706.0, "reward": 1.0220197439193726, "reward_std": 0.49257707595825195, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5440375208854675, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 2290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 467.5, "completions/mean_terminated_length": 467.5, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.4582, "grad_norm": 0.12495874613523483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26807834.0, "reward": 1.3180451393127441, "reward_std": 0.38987642526626587, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8235875368118286, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "step": 2291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 497.625, "completions/mean_terminated_length": 497.625, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.4584, "grad_norm": 0.11509034782648087, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26819716.0, "reward": 0.8763790130615234, "reward_std": 0.40148797631263733, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5027562379837036, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 2292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 559.75, "completions/mean_terminated_length": 559.75, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.4586, "grad_norm": 0.05864287540316582, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26833328.0, "reward": 0.692966639995575, "reward_std": 0.2821386754512787, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.32343125343322754, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.799375057220459, "step": 2293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 666.75, "completions/mean_terminated_length": 615.7142944335938, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.4588, "grad_norm": 0.11444199085235596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26848060.0, "reward": 0.8091410994529724, "reward_std": 0.42269131541252136, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6182812452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.46562498807907104, "step": 2294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 707.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 463.5625, "completions/mean_terminated_length": 463.5625, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.459, "grad_norm": 0.10244733095169067, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26859061.0, "reward": 1.2613259553909302, "reward_std": 0.4248977303504944, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7726500034332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 960.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 599.3125, "completions/mean_terminated_length": 599.3125, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.4592, "grad_norm": 0.073130302131176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26872618.0, "reward": 0.6983600854873657, "reward_std": 0.18035508692264557, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.39671874046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7156250476837158, "step": 2296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 635.1875, "completions/mean_terminated_length": 635.1875, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.4594, "grad_norm": 0.08073810487985611, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26886429.0, "reward": 0.8890600800514221, "reward_std": 0.354780375957489, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5281187295913696, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6931250095367432, "step": 2297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 888.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 326.25, "completions/mean_terminated_length": 326.25, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.4596, "grad_norm": 0.16658516228199005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26895041.0, "reward": 1.2877167463302612, "reward_std": 0.3898802101612091, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.825431227684021, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 2298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 389.875, "completions/mean_terminated_length": 389.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.4598, "grad_norm": 0.15161500871181488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26905727.0, "reward": 0.8979228138923645, "reward_std": 0.48130160570144653, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4208437502384186, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631250500679016, "step": 2299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 464.375, "completions/mean_terminated_length": 464.375, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.46, "grad_norm": 0.09299109876155853, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26917301.0, "reward": 0.6463509798049927, "reward_std": 0.33343350887298584, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16769999265670776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 2300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 509.1875, "completions/mean_terminated_length": 509.1875, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.4602, "grad_norm": 0.10328042507171631, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26929768.0, "reward": 0.7238291501998901, "reward_std": 0.3846589922904968, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2601562440395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 370.0, "completions/mean_terminated_length": 370.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.4604, "grad_norm": 0.12900826334953308, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26939336.0, "reward": 0.7242946028709412, "reward_std": 0.2675515115261078, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3860875070095062, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7649999856948853, "step": 2302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 696.9375, "completions/mean_terminated_length": 587.9166870117188, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.4606, "grad_norm": 0.10667295008897781, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26954775.0, "reward": 0.7630475759506226, "reward_std": 0.5882968306541443, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4635937511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6468750238418579, "step": 2303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 313.75, "completions/mean_terminated_length": 313.75, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.4608, "grad_norm": 0.2109200656414032, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26964371.0, "reward": 1.3208481073379517, "reward_std": 0.38387325406074524, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291937112808228, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 490.75, "completions/mean_terminated_length": 490.75, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.461, "grad_norm": 0.06652411818504333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26975999.0, "reward": 0.6014665961265564, "reward_std": 0.23970618844032288, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1404312551021576, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 663.75, "completions/mean_terminated_length": 612.2857666015625, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.4612, "grad_norm": 0.0779614970088005, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26990715.0, "reward": 0.8092848658561707, "reward_std": 0.3881639540195465, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6810687780380249, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3206250071525574, "step": 2306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 457.6875, "completions/mean_terminated_length": 457.6875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.4614, "grad_norm": 0.0569460391998291, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27002806.0, "reward": 0.7627350091934204, "reward_std": 0.15174254775047302, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5254687070846558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.640625, "step": 2307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 541.375, "completions/mean_terminated_length": 472.4285888671875, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.4616, "grad_norm": 0.29665279388427734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27015244.0, "reward": 1.0478134155273438, "reward_std": 0.5733710527420044, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6581249833106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 2308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 473.0, "completions/mean_terminated_length": 473.0, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.4618, "grad_norm": 0.10100781172513962, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27026636.0, "reward": 1.1041414737701416, "reward_std": 0.45143795013427734, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6457812786102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 2309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 380.6875, "completions/mean_terminated_length": 380.6875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.462, "grad_norm": 0.17694921791553497, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27036647.0, "reward": 1.1707634925842285, "reward_std": 0.5195708870887756, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7165249586105347, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 2310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 290.8125, "completions/mean_terminated_length": 290.8125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.4622, "grad_norm": 0.1620023399591446, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27045300.0, "reward": 1.3809137344360352, "reward_std": 0.3248433470726013, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8868250250816345, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 2311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 836.0, "completions/mean_terminated_length": 792.6154174804688, "completions/min_length": 518.0, "completions/min_terminated_length": 518.0, "epoch": 0.4624, "grad_norm": 0.08203538507223129, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27062532.0, "reward": 0.6900006532669067, "reward_std": 0.4599517285823822, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4424999952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5562499761581421, "step": 2312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 589.8125, "completions/mean_terminated_length": 589.8125, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.4626, "grad_norm": 0.15626260638237, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27077217.0, "reward": 0.9696851968765259, "reward_std": 0.5050912499427795, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.564368724822998, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8399999737739563, "step": 2313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 605.75, "completions/mean_terminated_length": 605.75, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.4628, "grad_norm": 0.07862674444913864, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27090717.0, "reward": 1.3769540786743164, "reward_std": 0.3329312205314636, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.87890625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 502.6875, "completions/mean_terminated_length": 502.6875, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.463, "grad_norm": 0.11880402266979218, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27102760.0, "reward": 1.0177886486053467, "reward_std": 0.49724605679512024, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5355750322341919, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 2315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 492.0, "completions/mean_terminated_length": 492.0, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.4632, "grad_norm": 0.15856090188026428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27114840.0, "reward": 1.0833414793014526, "reward_std": 0.48692676424980164, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6041812896728516, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 433.4375, "completions/mean_terminated_length": 433.4375, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.4634, "grad_norm": 0.07321865856647491, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27125503.0, "reward": 1.3892196416854858, "reward_std": 0.2039211541414261, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9659374952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7437499761581421, "step": 2317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 491.0, "completions/max_terminated_length": 491.0, "completions/mean_length": 396.875, "completions/mean_terminated_length": 396.875, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.4636, "grad_norm": 0.13503138720989227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27135469.0, "reward": 1.0101165771484375, "reward_std": 0.5054212212562561, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5202312469482422, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 2318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 235.0, "completions/max_terminated_length": 235.0, "completions/mean_length": 196.0, "completions/mean_terminated_length": 196.0, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.4638, "grad_norm": 0.0004997908254154027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27142381.0, "reward": 1.4991822242736816, "reward_std": 0.0005767570110037923, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983624815940857, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 2319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 654.5, "completions/mean_terminated_length": 629.86669921875, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.464, "grad_norm": 0.043366625905036926, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27156485.0, "reward": 0.5949977040290833, "reward_std": 0.2015310376882553, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2524937391281128, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7868750095367432, "step": 2320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 197.1875, "completions/mean_terminated_length": 197.1875, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.4642, "grad_norm": 0.19898252189159393, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27163464.0, "reward": 1.3178887367248535, "reward_std": 0.3897971212863922, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8232750296592712, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 2321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 549.5625, "completions/mean_terminated_length": 549.5625, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.4644, "grad_norm": 0.10796105861663818, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27176945.0, "reward": 0.6691384315490723, "reward_std": 0.2902933955192566, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4007750153541565, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5900000333786011, "step": 2322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 360.5, "completions/mean_terminated_length": 360.5, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.4646, "grad_norm": 0.15876486897468567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27186985.0, "reward": 1.3207042217254639, "reward_std": 0.382996141910553, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.828906238079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 494.8125, "completions/mean_terminated_length": 494.8125, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.4648, "grad_norm": 0.19887575507164001, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27198614.0, "reward": 0.7220197916030884, "reward_std": 0.3856043815612793, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2565374970436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 2324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 496.5625, "completions/mean_terminated_length": 496.5625, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.465, "grad_norm": 0.21657924354076385, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27210223.0, "reward": 1.0111603736877441, "reward_std": 0.4624027907848358, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5848187208175659, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 2325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 338.25, "completions/mean_terminated_length": 338.25, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.4652, "grad_norm": 0.13330981135368347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27219603.0, "reward": 0.6630476713180542, "reward_std": 0.2181936651468277, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.38859373331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7531249523162842, "step": 2326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 764.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 464.125, "completions/mean_terminated_length": 464.125, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.4654, "grad_norm": 0.14378398656845093, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27230869.0, "reward": 1.1677227020263672, "reward_std": 0.44832319021224976, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7104437351226807, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8831250071525574, "step": 2327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 507.8125, "completions/mean_terminated_length": 507.8125, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.4656, "grad_norm": 0.10637062788009644, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27242754.0, "reward": 0.8349322080612183, "reward_std": 0.46294519305229187, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3573625087738037, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 2328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 656.6875, "completions/mean_terminated_length": 656.6875, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.4658, "grad_norm": 0.07609909772872925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27257597.0, "reward": 0.7789069414138794, "reward_std": 0.2449372559785843, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.49531251192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 2329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 301.8125, "completions/mean_terminated_length": 301.8125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.466, "grad_norm": 0.13065184652805328, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27266522.0, "reward": 1.3103103637695312, "reward_std": 0.3403570353984833, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8706187009811401, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8806250095367432, "step": 2330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 378.1875, "completions/mean_terminated_length": 378.1875, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4662, "grad_norm": 0.17017921805381775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27276813.0, "reward": 1.1336729526519775, "reward_std": 0.43642470240592957, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7048437595367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 2331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 558.625, "completions/mean_terminated_length": 558.625, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.4664, "grad_norm": 0.20476804673671722, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27289703.0, "reward": 0.789441704750061, "reward_std": 0.3830791711807251, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.39138126373291016, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 2332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 284.0625, "completions/mean_terminated_length": 284.0625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.4666, "grad_norm": 0.14445272088050842, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27298024.0, "reward": 1.438016653060913, "reward_std": 0.24485822021961212, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9385312795639038, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 649.5625, "completions/mean_terminated_length": 649.5625, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.4668, "grad_norm": 0.11767327785491943, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27312113.0, "reward": 1.1988290548324585, "reward_std": 0.43842849135398865, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.710156261920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 2334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 710.1875, "completions/mean_terminated_length": 710.1875, "completions/min_length": 622.0, "completions/min_terminated_length": 622.0, "epoch": 0.467, "grad_norm": 0.034140799194574356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27327524.0, "reward": 1.4569041728973389, "reward_std": 0.1698419153690338, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9763063192367554, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 294.875, "completions/mean_terminated_length": 294.875, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.4672, "grad_norm": 0.11200561374425888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27337122.0, "reward": 0.8873445987701416, "reward_std": 0.19657720625400543, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7746875286102295, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.26249998807907104, "step": 2336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 609.5, "completions/mean_terminated_length": 581.86669921875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.4674, "grad_norm": 0.11318344622850418, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27350602.0, "reward": 0.6555479168891907, "reward_std": 0.28045669198036194, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3735937476158142, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6031249761581421, "step": 2337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 493.125, "completions/mean_terminated_length": 457.7333679199219, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.4676, "grad_norm": 0.05091384798288345, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27362396.0, "reward": 0.5433446764945984, "reward_std": 0.17445088922977448, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14918749034404755, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8512500524520874, "step": 2338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 287.875, "completions/mean_terminated_length": 287.875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.4678, "grad_norm": 0.243128702044487, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27370922.0, "reward": 0.8971041440963745, "reward_std": 0.48157986998558044, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41920626163482666, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 525.25, "completions/mean_terminated_length": 492.0000305175781, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.468, "grad_norm": 0.11907977610826492, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27383182.0, "reward": 1.3142790794372559, "reward_std": 0.43650588393211365, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8785562515258789, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 2340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 359.125, "completions/mean_terminated_length": 359.125, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.4682, "grad_norm": 0.07462160289287567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27392768.0, "reward": 1.39695405960083, "reward_std": 0.2825361490249634, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9189062714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 2341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 386.0625, "completions/mean_terminated_length": 386.0625, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.4684, "grad_norm": 0.10240267217159271, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27402897.0, "reward": 1.2613259553909302, "reward_std": 0.4248977303504944, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7726500034332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 674.125, "completions/mean_terminated_length": 650.800048828125, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.4686, "grad_norm": 0.16513358056545258, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27417331.0, "reward": 1.014375925064087, "reward_std": 0.5326734185218811, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5912500023841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "step": 2343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 648.1875, "completions/mean_terminated_length": 594.5, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.4688, "grad_norm": 0.09260932356119156, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27432438.0, "reward": 0.5996759533882141, "reward_std": 0.3981431722640991, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1993499994277954, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8299999833106995, "step": 2344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 507.25, "completions/mean_terminated_length": 507.25, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.469, "grad_norm": 0.08943504840135574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27444458.0, "reward": 1.3467978239059448, "reward_std": 0.33197957277297974, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8810937404632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 2345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 437.4375, "completions/mean_terminated_length": 437.4375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.4692, "grad_norm": 0.11557212471961975, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27455473.0, "reward": 1.2595164775848389, "reward_std": 0.4281834065914154, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.769031286239624, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 660.75, "completions/mean_terminated_length": 636.5333862304688, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.4694, "grad_norm": 0.08241716772317886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27470221.0, "reward": 0.925703763961792, "reward_std": 0.38980695605278015, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7264062762260437, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.46562498807907104, "step": 2347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 742.6875, "completions/mean_terminated_length": 723.933349609375, "completions/min_length": 566.0, "completions/min_terminated_length": 566.0, "epoch": 0.4696, "grad_norm": 0.1835465282201767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27486392.0, "reward": 1.1688883304595947, "reward_std": 0.5218823552131653, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7127749919891357, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 2348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 516.6875, "completions/mean_terminated_length": 516.6875, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.4698, "grad_norm": 0.1139196902513504, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27499811.0, "reward": 1.2029571533203125, "reward_std": 0.45333608984947205, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7184125185012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 534.875, "completions/mean_terminated_length": 534.875, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.47, "grad_norm": 0.0033854942303150892, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27513233.0, "reward": 0.551641583442688, "reward_std": 0.011562498286366463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10328124463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 455.9375, "completions/mean_terminated_length": 418.0666809082031, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.4702, "grad_norm": 0.1213795617222786, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27524688.0, "reward": 1.1641383171081543, "reward_std": 0.4946269989013672, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7657749652862549, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7900000214576721, "step": 2351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 483.9375, "completions/mean_terminated_length": 483.9375, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.4704, "grad_norm": 0.16954083740711212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27536559.0, "reward": 1.1253008842468262, "reward_std": 0.5102624893188477, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6880999803543091, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7987499833106995, "step": 2352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 531.0, "completions/mean_terminated_length": 531.0, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.4706, "grad_norm": 0.07758110016584396, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27549503.0, "reward": 0.6675007343292236, "reward_std": 0.16845716536045074, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3349999785423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.762499988079071, "step": 2353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 444.5625, "completions/mean_terminated_length": 444.5625, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.4708, "grad_norm": 0.2527989447116852, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27560344.0, "reward": 0.8438822031021118, "reward_std": 0.4561940133571625, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3752624988555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 852.3125, "completions/mean_terminated_length": 774.2727661132812, "completions/min_length": 487.0, "completions/min_terminated_length": 487.0, "epoch": 0.471, "grad_norm": 0.09087012708187103, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27577789.0, "reward": 0.522188127040863, "reward_std": 0.4767994284629822, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2318750023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6187499761581421, "step": 2355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 180.5625, "completions/mean_terminated_length": 180.5625, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.4712, "grad_norm": 0.10442069172859192, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27584166.0, "reward": 1.4399887323379517, "reward_std": 0.23766417801380157, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424749612808228, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 2356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 665.9375, "completions/mean_terminated_length": 665.9375, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.4714, "grad_norm": 0.07895689457654953, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27599493.0, "reward": 0.812410295009613, "reward_std": 0.4227410852909088, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37481874227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081250429153442, "step": 2357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 946.9375, "completions/mean_terminated_length": 869.875, "completions/min_length": 751.0, "completions/min_terminated_length": 751.0, "epoch": 0.4716, "grad_norm": 0.10877935588359833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27618756.0, "reward": 0.3983567953109741, "reward_std": 0.5082001090049744, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1717124879360199, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4712499976158142, "step": 2358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 726.3125, "completions/mean_terminated_length": 657.6154174804688, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.4718, "grad_norm": 0.05114375054836273, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27634249.0, "reward": 0.519806981086731, "reward_std": 0.3620305061340332, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16461250185966492, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7450000047683716, "step": 2359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 553.125, "completions/mean_terminated_length": 553.125, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.472, "grad_norm": 0.022200634703040123, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27646971.0, "reward": 0.5731134414672852, "reward_std": 0.08248577266931534, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1462249904870987, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 2360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 793.75, "completions/mean_terminated_length": 760.857177734375, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.4722, "grad_norm": 0.04645993560552597, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27663479.0, "reward": 0.6008599996566772, "reward_std": 0.29731881618499756, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3267187476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6281249523162842, "step": 2361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 984.5625, "completions/mean_terminated_length": 866.25, "completions/min_length": 852.0, "completions/min_terminated_length": 852.0, "epoch": 0.4724, "grad_norm": 0.04982580617070198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27684608.0, "reward": 0.17953142523765564, "reward_std": 0.33376410603523254, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10906250029802322, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.17499999701976776, "step": 2362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 392.9375, "completions/mean_terminated_length": 392.9375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.4726, "grad_norm": 0.19055619835853577, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27694767.0, "reward": 0.7154040932655334, "reward_std": 0.3889150321483612, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2433062493801117, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 2363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 276.5625, "completions/mean_terminated_length": 276.5625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.4728, "grad_norm": 0.4315725564956665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27702888.0, "reward": 0.7387759685516357, "reward_std": 0.4722534418106079, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2900499999523163, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 2364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 780.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 526.4375, "completions/mean_terminated_length": 526.4375, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.473, "grad_norm": 0.12472334504127502, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27715007.0, "reward": 0.820154070854187, "reward_std": 0.41967442631721497, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3903062641620636, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 2365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 342.4375, "completions/mean_terminated_length": 342.4375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4732, "grad_norm": 0.06321939080953598, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27725126.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 560.375, "completions/mean_terminated_length": 529.4666748046875, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.4734, "grad_norm": 0.10451162606477737, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27737628.0, "reward": 1.160601019859314, "reward_std": 0.5324415564537048, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6962000131607056, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.908750057220459, "step": 2367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 939.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 660.8125, "completions/mean_terminated_length": 660.8125, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.4736, "grad_norm": 0.22461135685443878, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27752729.0, "reward": 1.0486726760864258, "reward_std": 0.5604568719863892, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6598437428474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7406249642372131, "step": 2368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 445.875, "completions/mean_terminated_length": 445.875, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.4738, "grad_norm": 0.09712269902229309, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27764839.0, "reward": 0.9666134119033813, "reward_std": 0.4852484464645386, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49572503566741943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 538.0, "completions/mean_terminated_length": 505.60003662109375, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.474, "grad_norm": 0.07069984823465347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27777623.0, "reward": 0.8442972898483276, "reward_std": 0.26778480410575867, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7510937452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 2370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 471.0, "completions/max_terminated_length": 471.0, "completions/mean_length": 340.875, "completions/mean_terminated_length": 340.875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.4742, "grad_norm": 0.1808495968580246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27786805.0, "reward": 0.8293544054031372, "reward_std": 0.4668404757976532, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3462062478065491, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 2371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 247.375, "completions/mean_terminated_length": 247.375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.4744, "grad_norm": 0.09265913814306259, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27794299.0, "reward": 0.599685549736023, "reward_std": 0.2407803237438202, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13686874508857727, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 661.1875, "completions/mean_terminated_length": 637.0000610351562, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.4746, "grad_norm": 0.018449675291776657, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27809566.0, "reward": 0.51734459400177, "reward_std": 0.1384401023387909, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 2373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 392.75, "completions/mean_terminated_length": 392.75, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.4748, "grad_norm": 0.11811763793230057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27819354.0, "reward": 1.2522263526916504, "reward_std": 0.44263729453086853, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7544499635696411, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 2374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 284.75, "completions/mean_terminated_length": 284.75, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.475, "grad_norm": 0.1446947455406189, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27828246.0, "reward": 0.6596478223800659, "reward_std": 0.3278464674949646, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19429373741149902, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 486.5, "completions/mean_terminated_length": 486.5, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.4752, "grad_norm": 0.08368725329637527, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27840734.0, "reward": 0.6765384674072266, "reward_std": 0.32463619112968445, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22807498276233673, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9350000023841858, "step": 2376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 524.375, "completions/mean_terminated_length": 524.375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.4754, "grad_norm": 0.10889192670583725, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27853188.0, "reward": 1.1911571025848389, "reward_std": 0.40514516830444336, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.757312536239624, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8199999928474426, "step": 2377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 784.875, "completions/mean_terminated_length": 768.933349609375, "completions/min_length": 597.0, "completions/min_terminated_length": 597.0, "epoch": 0.4756, "grad_norm": 0.037978511303663254, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27870642.0, "reward": 0.5696883201599121, "reward_std": 0.19583453238010406, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2018750011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8125, "step": 2378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 448.875, "completions/mean_terminated_length": 448.875, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.4758, "grad_norm": 0.05237218737602234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27882128.0, "reward": 1.4083197116851807, "reward_std": 0.2614610195159912, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9416375160217285, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 2379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 496.25, "completions/mean_terminated_length": 496.25, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.476, "grad_norm": 0.09452932327985764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27893972.0, "reward": 0.6474226117134094, "reward_std": 0.2215365171432495, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.35734373331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.721875011920929, "step": 2380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 321.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 250.125, "completions/mean_terminated_length": 250.125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.4762, "grad_norm": 0.18265657126903534, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27901446.0, "reward": 1.321254014968872, "reward_std": 0.38326922059059143, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8300062417984009, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 2381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 484.4375, "completions/mean_terminated_length": 484.4375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.4764, "grad_norm": 0.09587187319993973, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27913773.0, "reward": 1.3190511465072632, "reward_std": 0.3870120942592621, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8256000280380249, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 2382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 419.125, "completions/mean_terminated_length": 419.125, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.4766, "grad_norm": 0.06794146448373795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27924399.0, "reward": 0.6303914785385132, "reward_std": 0.16826458275318146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26078125834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 2383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 704.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 557.75, "completions/mean_terminated_length": 557.75, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.4768, "grad_norm": 0.09240111708641052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27937323.0, "reward": 1.304335117340088, "reward_std": 0.3005831837654114, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.9211688041687012, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7268749475479126, "step": 2384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 500.6875, "completions/mean_terminated_length": 500.6875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.477, "grad_norm": 0.10915613919496536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27948870.0, "reward": 0.9207824468612671, "reward_std": 0.4768999218940735, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4665624797344208, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 2385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 344.0, "completions/mean_terminated_length": 344.0, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.4772, "grad_norm": 0.1971285194158554, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27958182.0, "reward": 1.0139323472976685, "reward_std": 0.5020154118537903, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.527862548828125, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 2386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 265.8125, "completions/mean_terminated_length": 265.8125, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.4774, "grad_norm": 0.18427826464176178, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27966723.0, "reward": 0.9020978808403015, "reward_std": 0.4777785539627075, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42919376492500305, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 417.0, "completions/mean_terminated_length": 417.0, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.4776, "grad_norm": 0.1626884639263153, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27977315.0, "reward": 1.3188165426254272, "reward_std": 0.38689541816711426, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8251312971115112, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 441.0666809082031, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.4778, "grad_norm": 0.271066814661026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27989179.0, "reward": 0.912804126739502, "reward_std": 0.5485377311706543, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5756062269210815, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7475000619888306, "step": 2389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 941.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 496.0625, "completions/mean_terminated_length": 496.0625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.478, "grad_norm": 0.10070188343524933, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28000812.0, "reward": 1.2613258361816406, "reward_std": 0.4248977601528168, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7726500034332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1022.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 637.5625, "completions/mean_terminated_length": 637.5625, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.4782, "grad_norm": 0.009037439711391926, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28015413.0, "reward": 0.9764846563339233, "reward_std": 0.0305000152438879, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9529687166213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.18437500298023224, "step": 2391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 409.8125, "completions/mean_terminated_length": 409.8125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.4784, "grad_norm": 0.005663058254867792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28025586.0, "reward": 0.5498321652412415, "reward_std": 0.014043880626559258, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09966249763965607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 423.875, "completions/mean_terminated_length": 423.875, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.4786, "grad_norm": 0.12283829599618912, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28035936.0, "reward": 1.0829694271087646, "reward_std": 0.3506108522415161, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7909374833106995, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 2393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 366.0, "completions/max_terminated_length": 366.0, "completions/mean_length": 227.4375, "completions/mean_terminated_length": 227.4375, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.4788, "grad_norm": 0.3062152862548828, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28043319.0, "reward": 1.1331791877746582, "reward_std": 0.48837319016456604, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6413562297821045, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 2394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 342.75, "completions/mean_terminated_length": 342.75, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.479, "grad_norm": 0.0002204361226176843, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28052627.0, "reward": 1.4991042613983154, "reward_std": 0.0005441238754428923, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982062578201294, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 2395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 889.625, "completions/mean_terminated_length": 809.0, "completions/min_length": 594.0, "completions/min_terminated_length": 594.0, "epoch": 0.4792, "grad_norm": 0.11578492820262909, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28070765.0, "reward": 0.4881255626678467, "reward_std": 0.4917002320289612, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22624999284744263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5625, "step": 2396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 918.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 661.4375, "completions/mean_terminated_length": 661.4375, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.4794, "grad_norm": 0.10867999494075775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28085540.0, "reward": 1.0146887302398682, "reward_std": 0.5003958344459534, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5293750166893005, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 2397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1018.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 559.4375, "completions/mean_terminated_length": 559.4375, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.4796, "grad_norm": 0.12163126468658447, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28098235.0, "reward": 0.9661728143692017, "reward_std": 0.43528392910957336, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5573437213897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 2398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 439.25, "completions/mean_terminated_length": 439.25, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.4798, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28109647.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 502.9375, "completions/mean_terminated_length": 502.9375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.48, "grad_norm": 0.1588299423456192, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28121886.0, "reward": 1.034454107284546, "reward_std": 0.4814393222332001, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5689062476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 466.375, "completions/mean_terminated_length": 466.375, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.4802, "grad_norm": 0.13681483268737793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28134036.0, "reward": 1.142632246017456, "reward_std": 0.4751051068305969, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6602624654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 318.625, "completions/mean_terminated_length": 318.625, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.4804, "grad_norm": 0.11064023524522781, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28143118.0, "reward": 1.3800790309906006, "reward_std": 0.32451745867729187, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851562738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 615.5625, "completions/mean_terminated_length": 615.5625, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.4806, "grad_norm": 0.05685434862971306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28157367.0, "reward": 0.5838297009468079, "reward_std": 0.24552008509635925, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10515624284744263, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781249761581421, "step": 2403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 457.3125, "completions/mean_terminated_length": 457.3125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.4808, "grad_norm": 0.05572507530450821, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28168748.0, "reward": 1.4398229122161865, "reward_std": 0.23761995136737823, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421437382698059, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 2404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 453.875, "completions/mean_terminated_length": 453.875, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.481, "grad_norm": 0.13871614634990692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28180186.0, "reward": 1.0242822170257568, "reward_std": 0.4904121458530426, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5485625267028809, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 2405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 793.4375, "completions/mean_terminated_length": 760.5000610351562, "completions/min_length": 443.0, "completions/min_terminated_length": 443.0, "epoch": 0.4812, "grad_norm": 0.06774145364761353, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28196721.0, "reward": 0.45530104637145996, "reward_std": 0.24194499850273132, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09809999912977219, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8174999952316284, "step": 2406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 238.0, "completions/max_terminated_length": 238.0, "completions/mean_length": 189.75, "completions/mean_terminated_length": 189.75, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.4814, "grad_norm": 0.3878304660320282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28203421.0, "reward": 1.0168511867523193, "reward_std": 0.4986940920352936, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5336999893188477, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 2407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 359.9375, "completions/mean_terminated_length": 359.9375, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.4816, "grad_norm": 0.056908994913101196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28213644.0, "reward": 1.4676573276519775, "reward_std": 0.12470872700214386, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9978125095367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 2408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 568.0625, "completions/mean_terminated_length": 568.0625, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.4818, "grad_norm": 0.14760799705982208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28226429.0, "reward": 0.9802230596542358, "reward_std": 0.48660919070243835, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5229437947273254, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143749475479126, "step": 2409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 699.625, "completions/mean_terminated_length": 591.5, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.482, "grad_norm": 0.14532342553138733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28241847.0, "reward": 0.6661725044250488, "reward_std": 0.5291886329650879, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3948437571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5718749761581421, "step": 2410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 556.0, "completions/mean_terminated_length": 524.800048828125, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.4822, "grad_norm": 0.12576737999916077, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28254295.0, "reward": 0.6580477952957153, "reward_std": 0.3756336569786072, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2535937428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 2411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 287.125, "completions/mean_terminated_length": 287.125, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.4824, "grad_norm": 0.0003891945234499872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28262825.0, "reward": 1.4994041919708252, "reward_std": 0.0005983232986181974, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9988062381744385, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 2412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 524.8125, "completions/mean_terminated_length": 524.8125, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.4826, "grad_norm": 0.06451676040887833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28276710.0, "reward": 0.8067195415496826, "reward_std": 0.23867161571979523, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5509375333786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 2413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 489.9375, "completions/mean_terminated_length": 489.9375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.4828, "grad_norm": 0.08449845016002655, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28288677.0, "reward": 0.6894540786743164, "reward_std": 0.2778562009334564, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.796875, "step": 2414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 484.125, "completions/mean_terminated_length": 484.125, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.483, "grad_norm": 0.026997817680239677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28300791.0, "reward": 0.5761728286743164, "reward_std": 0.07956567406654358, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15234375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 2415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 377.875, "completions/mean_terminated_length": 377.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.4832, "grad_norm": 0.09276429563760757, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28310741.0, "reward": 1.3808326721191406, "reward_std": 0.32481181621551514, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.886662483215332, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "step": 2416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 833.0625, "completions/mean_terminated_length": 746.2727661132812, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.4834, "grad_norm": 0.11457730084657669, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28327894.0, "reward": 0.4388037621974945, "reward_std": 0.38476642966270447, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12760624289512634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6506249904632568, "step": 2417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 484.0, "completions/mean_terminated_length": 484.0, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.4836, "grad_norm": 0.11426842957735062, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28339622.0, "reward": 1.1918261051177979, "reward_std": 0.47088098526000977, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6961500644683838, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 440.1875, "completions/mean_terminated_length": 401.2666931152344, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.4838, "grad_norm": 0.1426951140165329, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28351001.0, "reward": 0.9538851976394653, "reward_std": 0.5290578007698059, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.532768726348877, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.846875011920929, "step": 2419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 769.0, "completions/mean_terminated_length": 710.1538696289062, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.484, "grad_norm": 0.11397615075111389, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28367001.0, "reward": 0.49093860387802124, "reward_std": 0.2936798930168152, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1693750023841858, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6749999523162842, "step": 2420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 356.6875, "completions/mean_terminated_length": 356.6875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.4842, "grad_norm": 0.0639968290925026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28376756.0, "reward": 1.4366167783737183, "reward_std": 0.24976444244384766, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9357312321662903, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 494.9375, "completions/mean_terminated_length": 494.9375, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.4844, "grad_norm": 0.0923265814781189, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28388163.0, "reward": 0.7326571941375732, "reward_std": 0.3804062604904175, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27781248092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 448.5, "completions/mean_terminated_length": 448.5, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.4846, "grad_norm": 0.05596695467829704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28399243.0, "reward": 0.5931978225708008, "reward_std": 0.1471492350101471, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18639373779296875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 2423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 610.25, "completions/mean_terminated_length": 582.6666870117188, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.4848, "grad_norm": 0.05414256080985069, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28413487.0, "reward": 0.6142196655273438, "reward_std": 0.23981218039989471, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.29093748331069946, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 2424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 201.0, "completions/max_terminated_length": 201.0, "completions/mean_length": 149.4375, "completions/mean_terminated_length": 149.4375, "completions/min_length": 107.0, "completions/min_terminated_length": 107.0, "epoch": 0.485, "grad_norm": 0.009872116148471832, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28419574.0, "reward": 0.5178415775299072, "reward_std": 0.015528772957623005, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03568124771118164, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 2425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 991.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 556.9375, "completions/mean_terminated_length": 556.9375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.4852, "grad_norm": 0.03360613435506821, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28432645.0, "reward": 0.6160165071487427, "reward_std": 0.10703200846910477, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23203124105930328, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656250238418579, "step": 2426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 495.625, "completions/mean_terminated_length": 495.625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.4854, "grad_norm": 0.11950092017650604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28444367.0, "reward": 0.7519540786743164, "reward_std": 0.46550774574279785, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 534.1875, "completions/mean_terminated_length": 370.91668701171875, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.4856, "grad_norm": 0.1287718117237091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28457170.0, "reward": 0.5211727619171143, "reward_std": 0.44754672050476074, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16734375059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.721875011920929, "step": 2428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 612.8125, "completions/mean_terminated_length": 585.4000244140625, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.4858, "grad_norm": 0.04794713109731674, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28470783.0, "reward": 0.9795316457748413, "reward_std": 0.318645715713501, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8965625166893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.25, "step": 2429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 812.1875, "completions/mean_terminated_length": 685.1000366210938, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.486, "grad_norm": 0.08528465032577515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28487570.0, "reward": 0.399919331073761, "reward_std": 0.39634644985198975, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11233749985694885, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5962499976158142, "step": 2430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 397.875, "completions/mean_terminated_length": 397.875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.4862, "grad_norm": 0.07532036304473877, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28497664.0, "reward": 0.6332696080207825, "reward_std": 0.2549535036087036, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20403750240802765, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 2431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 593.125, "completions/mean_terminated_length": 593.125, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.4864, "grad_norm": 0.08461067825555801, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28511170.0, "reward": 1.1373885869979858, "reward_std": 0.4827633500099182, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.649774968624115, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 704.25, "completions/mean_terminated_length": 630.4615478515625, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.4866, "grad_norm": 0.07846013456583023, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28527510.0, "reward": 0.4249195158481598, "reward_std": 0.25888803601264954, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09983749687671661, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7556250095367432, "step": 2433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 956.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 544.5625, "completions/mean_terminated_length": 544.5625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.4868, "grad_norm": 0.1002664715051651, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28539951.0, "reward": 0.5832822322845459, "reward_std": 0.22088722884655, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2290624976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 2434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 476.625, "completions/mean_terminated_length": 440.13336181640625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.487, "grad_norm": 0.11971651762723923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28552057.0, "reward": 0.9700758457183838, "reward_std": 0.45358625054359436, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6901500225067139, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5462499856948853, "step": 2435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 494.875, "completions/mean_terminated_length": 494.875, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.4872, "grad_norm": 0.23619140684604645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28563623.0, "reward": 0.8779850602149963, "reward_std": 0.5154407620429993, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44346821308135986, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9416687488555908, "step": 2436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 471.5, "completions/mean_terminated_length": 471.5, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.4874, "grad_norm": 0.085592120885849, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28575711.0, "reward": 0.6291260719299316, "reward_std": 0.2594689726829529, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19574999809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 2437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 322.0, "completions/mean_terminated_length": 322.0, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.4876, "grad_norm": 9.714515908854082e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28584559.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 424.9375, "completions/mean_terminated_length": 424.9375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.4878, "grad_norm": 0.10612854361534119, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28596270.0, "reward": 0.6202228665351868, "reward_std": 0.2446165829896927, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17794375121593475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331249594688416, "step": 2439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 427.875, "completions/mean_terminated_length": 427.875, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.488, "grad_norm": 0.18076658248901367, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28607116.0, "reward": 1.0186604261398315, "reward_std": 0.49820607900619507, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.662318766117096, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7831249833106995, "step": 2440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 573.25, "completions/mean_terminated_length": 543.2000122070312, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.4882, "grad_norm": 0.08387967944145203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28619776.0, "reward": 0.6853134036064148, "reward_std": 0.36181843280792236, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30812501907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.831250011920929, "step": 2441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 791.8125, "completions/mean_terminated_length": 714.4166870117188, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.4884, "grad_norm": 0.07899794727563858, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28636189.0, "reward": 0.4144538342952728, "reward_std": 0.2473968267440796, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07890624552965164, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 2442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 540.4375, "completions/mean_terminated_length": 508.20001220703125, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.4886, "grad_norm": 0.12103099375963211, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28648660.0, "reward": 0.6935164928436279, "reward_std": 0.3704499304294586, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.262031227350235, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.796875, "step": 2443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 311.0, "completions/max_terminated_length": 311.0, "completions/mean_length": 233.625, "completions/mean_terminated_length": 233.625, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.4888, "grad_norm": 0.26622962951660156, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28656142.0, "reward": 0.9598071575164795, "reward_std": 0.4916360378265381, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4821125268936157, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 2444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 364.0625, "completions/mean_terminated_length": 364.0625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.489, "grad_norm": 0.20385773479938507, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28665903.0, "reward": 1.190794825553894, "reward_std": 0.4727404713630676, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6940875053405762, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 2445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 540.625, "completions/mean_terminated_length": 540.625, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.4892, "grad_norm": 0.0749635249376297, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28678521.0, "reward": 0.6746103763580322, "reward_std": 0.2563198208808899, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2867187261581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 2446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 648.5625, "completions/mean_terminated_length": 594.9285888671875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.4894, "grad_norm": 0.07829739898443222, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28692546.0, "reward": 0.5202351808547974, "reward_std": 0.2274371087551117, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1654687523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.778124988079071, "step": 2447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 477.25, "completions/mean_terminated_length": 440.8000183105469, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.4896, "grad_norm": 0.12159845978021622, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28704406.0, "reward": 1.0430355072021484, "reward_std": 0.5494503378868103, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5860687494277954, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 2448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 513.0625, "completions/mean_terminated_length": 479.0000305175781, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.4898, "grad_norm": 0.14925901591777802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28716247.0, "reward": 1.072188377380371, "reward_std": 0.5260415077209473, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6443749666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 2449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 787.25, "completions/mean_terminated_length": 771.4667358398438, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.49, "grad_norm": 0.07511844485998154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28732763.0, "reward": 0.7453913688659668, "reward_std": 0.37503311038017273, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4282812476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 2450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 294.0, "completions/mean_terminated_length": 294.0, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.4902, "grad_norm": 0.1704065501689911, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28741035.0, "reward": 0.6638821959495544, "reward_std": 0.32604628801345825, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20276248455047607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 351.0, "completions/max_terminated_length": 351.0, "completions/mean_length": 235.75, "completions/mean_terminated_length": 235.75, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.4904, "grad_norm": 0.07502123713493347, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28748967.0, "reward": 1.4371891021728516, "reward_std": 0.2499171644449234, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9368749856948853, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.987500011920929, "step": 2452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 707.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 567.6875, "completions/mean_terminated_length": 567.6875, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.4906, "grad_norm": 0.09319186210632324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28762514.0, "reward": 1.2553918361663818, "reward_std": 0.436075359582901, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7607812881469727, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 2453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 997.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 627.0, "completions/mean_terminated_length": 627.0, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.4908, "grad_norm": 0.10065745562314987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28776706.0, "reward": 1.0552351474761963, "reward_std": 0.3897286653518677, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7354687452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6843749284744263, "step": 2454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 666.8125, "completions/mean_terminated_length": 615.7857666015625, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.491, "grad_norm": 0.15206031501293182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28793215.0, "reward": 0.8348439931869507, "reward_std": 0.43327850103378296, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.7946875095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.19374999403953552, "step": 2455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 342.6875, "completions/mean_terminated_length": 342.6875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.4912, "grad_norm": 0.20487597584724426, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28802490.0, "reward": 1.0222822427749634, "reward_std": 0.4928482472896576, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5445625185966492, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 2456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 531.6875, "completions/mean_terminated_length": 531.6875, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.4914, "grad_norm": 0.07063426822423935, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28815605.0, "reward": 0.8589849472045898, "reward_std": 0.18741784989833832, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.717968761920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4156250059604645, "step": 2457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 440.4375, "completions/mean_terminated_length": 440.4375, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.4916, "grad_norm": 0.1187933012843132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28827180.0, "reward": 0.9631103277206421, "reward_std": 0.4883483052253723, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48871874809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 413.375, "completions/mean_terminated_length": 413.375, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.4918, "grad_norm": 0.18465110659599304, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28837794.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 221.0, "completions/max_terminated_length": 221.0, "completions/mean_length": 176.4375, "completions/mean_terminated_length": 176.4375, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.492, "grad_norm": 0.3617279827594757, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28844073.0, "reward": 0.879957377910614, "reward_std": 0.4960055649280548, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38491249084472656, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9912500381469727, "step": 2460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 497.8125, "completions/mean_terminated_length": 376.3846435546875, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 0.4922, "grad_norm": 0.2300676703453064, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28855830.0, "reward": 0.6382038593292236, "reward_std": 0.46050742268562317, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.46390625834465027, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3843749761581421, "step": 2461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 299.875, "completions/mean_terminated_length": 299.875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.4924, "grad_norm": 0.23775525391101837, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28864548.0, "reward": 1.2865259647369385, "reward_std": 0.47121289372444153, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8230500221252441, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 448.25, "completions/mean_terminated_length": 448.25, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.4926, "grad_norm": 0.10450060665607452, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28876280.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 815.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 446.3125, "completions/mean_terminated_length": 446.3125, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.4928, "grad_norm": 0.17240199446678162, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28887309.0, "reward": 0.9677321314811707, "reward_std": 0.4840494990348816, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4979625344276428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 2464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 487.0625, "completions/mean_terminated_length": 451.2666931152344, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.493, "grad_norm": 0.06154828146100044, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28898894.0, "reward": 0.5720195770263672, "reward_std": 0.28234758973121643, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14403750002384186, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 2465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 498.0, "completions/mean_terminated_length": 462.933349609375, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.4932, "grad_norm": 0.2034861445426941, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28910862.0, "reward": 0.7315633893013, "reward_std": 0.4240562319755554, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.33812499046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 2466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 604.5625, "completions/mean_terminated_length": 544.6428833007812, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.4934, "grad_norm": 0.1549360454082489, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28924471.0, "reward": 1.0674231052398682, "reward_std": 0.595824658870697, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6348437070846558, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.840624988079071, "step": 2467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 447.1875, "completions/mean_terminated_length": 447.1875, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.4936, "grad_norm": 0.0996013879776001, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28935642.0, "reward": 1.377941608428955, "reward_std": 0.3313320279121399, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8808813095092773, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 686.875, "completions/mean_terminated_length": 533.6363525390625, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.4938, "grad_norm": 0.16925814747810364, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28950328.0, "reward": 0.6207820177078247, "reward_std": 0.5798870921134949, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30406248569488525, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.65625, "step": 2469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 352.0, "completions/max_terminated_length": 352.0, "completions/mean_length": 304.4375, "completions/mean_terminated_length": 304.4375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.494, "grad_norm": 0.00026406548568047583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28958991.0, "reward": 1.4992698431015015, "reward_std": 0.0006088929949328303, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998537540435791, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 504.125, "completions/mean_terminated_length": 504.125, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.4942, "grad_norm": 0.10022828727960587, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28971233.0, "reward": 1.2332819700241089, "reward_std": 0.34975653886795044, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8415625095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 2471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 864.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 484.625, "completions/mean_terminated_length": 484.625, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.4944, "grad_norm": 0.13123741745948792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28982715.0, "reward": 1.1018508672714233, "reward_std": 0.4665924906730652, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6412000060081482, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9162499904632568, "step": 2472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 465.875, "completions/mean_terminated_length": 465.875, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.4946, "grad_norm": 0.21782030165195465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28993913.0, "reward": 0.820385217666626, "reward_std": 0.37977132201194763, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.45326876640319824, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7831249833106995, "step": 2473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 440.375, "completions/mean_terminated_length": 440.375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.4948, "grad_norm": 0.17064914107322693, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29004815.0, "reward": 1.0182760953903198, "reward_std": 0.4970442056655884, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.536549985408783, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 250.0625, "completions/mean_terminated_length": 250.0625, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.495, "grad_norm": 0.009050389751791954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29012960.0, "reward": 0.5245009660720825, "reward_std": 0.01939997635781765, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04899999871850014, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 643.375, "completions/mean_terminated_length": 555.5384521484375, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.4952, "grad_norm": 0.09299473464488983, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29027814.0, "reward": 0.527966320514679, "reward_std": 0.3059108257293701, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2434312403202057, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6368749737739563, "step": 2476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 504.375, "completions/mean_terminated_length": 504.375, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.4954, "grad_norm": 0.11003368347883224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29040892.0, "reward": 0.7832041382789612, "reward_std": 0.42683979868888855, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 637.4375, "completions/mean_terminated_length": 582.2142944335938, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.4956, "grad_norm": 0.051496218889951706, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29055443.0, "reward": 0.4908353090286255, "reward_std": 0.2099929302930832, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10666874051094055, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8131250143051147, "step": 2478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 940.0625, "completions/mean_terminated_length": 832.1428833007812, "completions/min_length": 678.0, "completions/min_terminated_length": 678.0, "epoch": 0.4958, "grad_norm": 0.08647966384887695, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29074836.0, "reward": 0.3892970681190491, "reward_std": 0.4669044613838196, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3410937488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.15937501192092896, "step": 2479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 600.75, "completions/mean_terminated_length": 600.75, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.496, "grad_norm": 0.006115803960710764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29088720.0, "reward": 0.5309855937957764, "reward_std": 0.02742534503340721, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06196874752640724, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 2480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 349.5, "completions/mean_terminated_length": 349.5, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.4962, "grad_norm": 0.21011941134929657, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29098296.0, "reward": 0.9602822065353394, "reward_std": 0.49087631702423096, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4830625057220459, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 735.625, "completions/mean_terminated_length": 716.4000244140625, "completions/min_length": 544.0, "completions/min_terminated_length": 544.0, "epoch": 0.4964, "grad_norm": 0.03867612034082413, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29113906.0, "reward": 0.5442976951599121, "reward_std": 0.16671845316886902, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1510937362909317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 2482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 370.0, "completions/mean_terminated_length": 370.0, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.4966, "grad_norm": 0.21404381096363068, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29124082.0, "reward": 0.9700665473937988, "reward_std": 0.4812568128108978, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5026312470436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 2483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 404.8125, "completions/mean_terminated_length": 404.8125, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.4968, "grad_norm": 0.1862718015909195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29134431.0, "reward": 0.7534383535385132, "reward_std": 0.3494427800178528, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31937500834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 2484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 712.8125, "completions/mean_terminated_length": 668.357177734375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.497, "grad_norm": 0.10266333073377609, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29150204.0, "reward": 0.6827352046966553, "reward_std": 0.452667772769928, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.30296874046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 2485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 431.625, "completions/mean_terminated_length": 431.625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.4972, "grad_norm": 0.13777148723602295, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29161414.0, "reward": 0.8438165187835693, "reward_std": 0.456093430519104, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3751312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 663.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 496.9375, "completions/mean_terminated_length": 496.9375, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.4974, "grad_norm": 0.12413391470909119, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29176053.0, "reward": 0.9619508981704712, "reward_std": 0.4890301525592804, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4864000082015991, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 539.8125, "completions/mean_terminated_length": 507.5333557128906, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.4976, "grad_norm": 0.1038459986448288, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29188546.0, "reward": 0.8290630578994751, "reward_std": 0.3396240770816803, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5956249833106995, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.574999988079071, "step": 2488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 519.5, "completions/mean_terminated_length": 519.5, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.4978, "grad_norm": 0.08513244241476059, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29200842.0, "reward": 1.315204381942749, "reward_std": 0.3519928455352783, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8804062008857727, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 2489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 371.125, "completions/mean_terminated_length": 371.125, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.498, "grad_norm": 0.166996031999588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29211244.0, "reward": 1.0832791328430176, "reward_std": 0.4868698716163635, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6040562391281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 869.0, "completions/max_terminated_length": 869.0, "completions/mean_length": 424.9375, "completions/mean_terminated_length": 424.9375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.4982, "grad_norm": 0.00788206048309803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29221899.0, "reward": 0.5413135886192322, "reward_std": 0.023594455793499947, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08262499421834946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 880.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 406.875, "completions/mean_terminated_length": 406.875, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.4984, "grad_norm": 0.1695777326822281, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29232777.0, "reward": 0.9019417762756348, "reward_std": 0.4782165586948395, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4288812577724457, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 530.125, "completions/mean_terminated_length": 530.125, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.4986, "grad_norm": 0.002381535479798913, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29244811.0, "reward": 0.551641583442688, "reward_std": 0.011562498286366463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10328124463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 665.75, "completions/mean_terminated_length": 614.5714721679688, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.4988, "grad_norm": 0.06561622023582458, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29259463.0, "reward": 0.5839725732803345, "reward_std": 0.2791481018066406, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2929437458515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6706249713897705, "step": 2494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 430.25, "completions/mean_terminated_length": 430.25, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.499, "grad_norm": 0.09605423361063004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29270395.0, "reward": 1.3971104621887207, "reward_std": 0.2825965881347656, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9192187786102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 2495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 298.1875, "completions/mean_terminated_length": 298.1875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.4992, "grad_norm": 0.22581294178962708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29278894.0, "reward": 1.0166230201721191, "reward_std": 0.4984584450721741, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5332437753677368, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 846.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 554.875, "completions/mean_terminated_length": 554.875, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.4994, "grad_norm": 0.09325361251831055, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29291404.0, "reward": 1.1897668838500977, "reward_std": 0.47363564372062683, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6920312643051147, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 2497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 496.5625, "completions/mean_terminated_length": 496.5625, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.4996, "grad_norm": 0.09943579882383347, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29303637.0, "reward": 1.315091609954834, "reward_std": 0.39646923542022705, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8176810145378113, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9742500185966492, "step": 2498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 409.8125, "completions/mean_terminated_length": 409.8125, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.4998, "grad_norm": 0.04402089864015579, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29314834.0, "reward": 0.6191137433052063, "reward_std": 0.1535012573003769, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23822569847106934, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8426874876022339, "step": 2499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 729.0, "completions/mean_terminated_length": 686.857177734375, "completions/min_length": 522.0, "completions/min_terminated_length": 522.0, "epoch": 0.5, "grad_norm": 0.12337329238653183, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29330562.0, "reward": 0.8547537922859192, "reward_std": 0.5264956951141357, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5220062732696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6956250071525574, "step": 2500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 502.9375, "completions/mean_terminated_length": 502.9375, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.5002, "grad_norm": 0.128530353307724, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29342625.0, "reward": 1.1060947179794312, "reward_std": 0.46685487031936646, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6496874690055847, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9187499284744263, "step": 2501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 389.6875, "completions/mean_terminated_length": 389.6875, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.5004, "grad_norm": 0.00012867053737863898, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29352988.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 342.625, "completions/mean_terminated_length": 342.625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.5006, "grad_norm": 0.25073152780532837, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29362198.0, "reward": 1.0129605531692505, "reward_std": 0.5022997260093689, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.52591872215271, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 2503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 339.375, "completions/mean_terminated_length": 339.375, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.5008, "grad_norm": 0.09004200994968414, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29371324.0, "reward": 1.4423322677612305, "reward_std": 0.22595512866973877, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9471625089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 926.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 608.3125, "completions/mean_terminated_length": 608.3125, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.501, "grad_norm": 0.040459971874952316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29385617.0, "reward": 0.9388285279273987, "reward_std": 0.18594035506248474, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.8151562809944153, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.37812501192092896, "step": 2505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 573.0, "completions/mean_terminated_length": 573.0, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.5012, "grad_norm": 0.11472834646701813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29399009.0, "reward": 0.8891413807868958, "reward_std": 0.4109727740287781, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.46578121185302734, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 2506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 713.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 528.5, "completions/mean_terminated_length": 528.5, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.5014, "grad_norm": 0.052733950316905975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29411145.0, "reward": 0.6038259267807007, "reward_std": 0.23084959387779236, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14515000581741333, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "step": 2507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 574.125, "completions/mean_terminated_length": 424.16668701171875, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.5016, "grad_norm": 0.27068305015563965, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29423883.0, "reward": 0.8181259632110596, "reward_std": 0.6516907215118408, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.44874998927116394, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 2508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 448.25, "completions/mean_terminated_length": 448.25, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.5018, "grad_norm": 0.0688302218914032, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29435247.0, "reward": 0.6282822489738464, "reward_std": 0.2505326271057129, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1940625011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 2509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 570.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 374.0625, "completions/mean_terminated_length": 374.0625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.502, "grad_norm": 0.23376956582069397, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29444976.0, "reward": 0.9699884057044983, "reward_std": 0.4814980924129486, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5024750232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 2510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 856.5625, "completions/mean_terminated_length": 780.45458984375, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.5022, "grad_norm": 0.09128629416227341, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29463049.0, "reward": 0.5643755197525024, "reward_std": 0.4981890618801117, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3162499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5375000238418579, "step": 2511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 274.0, "completions/max_terminated_length": 274.0, "completions/mean_length": 234.375, "completions/mean_terminated_length": 234.375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.5024, "grad_norm": 0.1294354349374771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29471327.0, "reward": 1.4400511980056763, "reward_std": 0.23768079280853271, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9426000118255615, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 2512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 507.4375, "completions/mean_terminated_length": 433.64288330078125, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.5026, "grad_norm": 0.1550961285829544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29483030.0, "reward": 0.6985039114952087, "reward_std": 0.4261552393436432, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.39700624346733093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6643750071525574, "step": 2513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 483.6875, "completions/mean_terminated_length": 483.6875, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.5028, "grad_norm": 0.0766160860657692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29494689.0, "reward": 0.611016571521759, "reward_std": 0.23701010644435883, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15953123569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 440.3125, "completions/mean_terminated_length": 440.3125, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.503, "grad_norm": 0.24506153166294098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29506614.0, "reward": 1.1675790548324585, "reward_std": 0.5229056477546692, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.710156261920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 475.0625, "completions/mean_terminated_length": 438.4666748046875, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.5032, "grad_norm": 0.13029079139232635, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29518407.0, "reward": 0.7107788324356079, "reward_std": 0.38562726974487305, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3590562343597412, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7493749856948853, "step": 2516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 896.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 565.9375, "completions/mean_terminated_length": 565.9375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.5034, "grad_norm": 0.06659342348575592, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29532150.0, "reward": 0.6732820272445679, "reward_std": 0.18470992147922516, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3465625047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 2517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 302.0, "completions/max_terminated_length": 302.0, "completions/mean_length": 250.5, "completions/mean_terminated_length": 250.5, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.5036, "grad_norm": 0.00017029023729264736, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29540126.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 672.625, "completions/mean_terminated_length": 591.5385131835938, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.5038, "grad_norm": 0.08907780796289444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29555512.0, "reward": 0.5768570303916931, "reward_std": 0.3568952977657318, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.34121251106262207, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49125003814697266, "step": 2519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 669.3125, "completions/mean_terminated_length": 645.6666870117188, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.504, "grad_norm": 0.08186574280261993, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29570125.0, "reward": 0.736847996711731, "reward_std": 0.4731264114379883, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28619372844696045, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9093750715255737, "step": 2520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 608.6875, "completions/mean_terminated_length": 512.84619140625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.5042, "grad_norm": 0.13646310567855835, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29583752.0, "reward": 0.6947535872459412, "reward_std": 0.4201577305793762, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5145062208175659, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.46437501907348633, "step": 2521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 374.0, "completions/mean_terminated_length": 374.0, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.5044, "grad_norm": 0.1933378279209137, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29593912.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 407.5625, "completions/mean_terminated_length": 407.5625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.5046, "grad_norm": 0.14421160519123077, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29604417.0, "reward": 0.9025759100914001, "reward_std": 0.47703468799591064, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43015000224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 502.1875, "completions/mean_terminated_length": 502.1875, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.5048, "grad_norm": 0.09853158891201019, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29616612.0, "reward": 0.6918728351593018, "reward_std": 0.3335920572280884, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2587437629699707, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 2524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 615.125, "completions/mean_terminated_length": 615.125, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.505, "grad_norm": 0.08422977477312088, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29630358.0, "reward": 1.1939727067947388, "reward_std": 0.4169568121433258, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7629437446594238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8831249475479126, "step": 2525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 419.875, "completions/mean_terminated_length": 419.875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.5052, "grad_norm": 0.0002165433979826048, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29641188.0, "reward": 1.4991259574890137, "reward_std": 0.0005745142116211355, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982500076293945, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 2526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 312.4375, "completions/mean_terminated_length": 312.4375, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.5054, "grad_norm": 0.1627415418624878, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29650779.0, "reward": 0.7633854150772095, "reward_std": 0.4390362501144409, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2767687439918518, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 2527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 791.1875, "completions/mean_terminated_length": 685.3636474609375, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.5056, "grad_norm": 0.10356847941875458, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29669214.0, "reward": 0.46203184127807617, "reward_std": 0.4043804705142975, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17406249046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6124999523162842, "step": 2528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 370.625, "completions/mean_terminated_length": 370.625, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.5058, "grad_norm": 0.10653560608625412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29679064.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 233.125, "completions/mean_terminated_length": 233.125, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.506, "grad_norm": 0.26515495777130127, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29687146.0, "reward": 0.9242572784423828, "reward_std": 0.47249603271484375, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4735125005245209, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 2530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 753.8125, "completions/mean_terminated_length": 691.4615478515625, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.5062, "grad_norm": 0.12942220270633698, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29703671.0, "reward": 0.8103851079940796, "reward_std": 0.5919363498687744, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4332687258720398, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7643749713897705, "step": 2531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 751.9375, "completions/mean_terminated_length": 733.800048828125, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.5064, "grad_norm": 0.049954064190387726, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29720182.0, "reward": 0.6944537162780762, "reward_std": 0.25531917810440063, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4514062702655792, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6218750476837158, "step": 2532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 469.875, "completions/mean_terminated_length": 469.875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.5066, "grad_norm": 0.00823415257036686, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29731988.0, "reward": 0.5423449873924255, "reward_std": 0.029583048075437546, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0846875011920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 293.5625, "completions/mean_terminated_length": 293.5625, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.5068, "grad_norm": 0.1975695937871933, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29741149.0, "reward": 0.8683228492736816, "reward_std": 0.45367151498794556, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4241437613964081, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9118750095367432, "step": 2534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 524.3125, "completions/mean_terminated_length": 524.3125, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.507, "grad_norm": 0.1270352602005005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29753634.0, "reward": 1.317501187324524, "reward_std": 0.3896059989929199, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8224999904632568, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 506.4375, "completions/mean_terminated_length": 506.4375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.5072, "grad_norm": 0.09853904694318771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29765577.0, "reward": 1.0224199295043945, "reward_std": 0.45904940366744995, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.607337474822998, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8337500095367432, "step": 2536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 417.5, "completions/mean_terminated_length": 417.5, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.5074, "grad_norm": 0.08938639611005783, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29775873.0, "reward": 0.633747935295105, "reward_std": 0.257753849029541, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20499375462532043, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 2537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 415.4375, "completions/mean_terminated_length": 415.4375, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.5076, "grad_norm": 0.20395079255104065, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29786344.0, "reward": 1.020769715309143, "reward_std": 0.4939327836036682, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5415375232696533, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 398.3125, "completions/mean_terminated_length": 398.3125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.5078, "grad_norm": 0.13637655973434448, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29797613.0, "reward": 1.318960428237915, "reward_std": 0.3869669437408447, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8254187107086182, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 589.4375, "completions/mean_terminated_length": 589.4375, "completions/min_length": 454.0, "completions/min_terminated_length": 454.0, "epoch": 0.508, "grad_norm": 0.08542684465646744, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29810932.0, "reward": 0.7416417002677917, "reward_std": 0.3271328806877136, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3582812547683716, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 2540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 418.875, "completions/mean_terminated_length": 378.5333557128906, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5082, "grad_norm": 0.09317339211702347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29821570.0, "reward": 0.5738290548324585, "reward_std": 0.28208333253860474, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 365.375, "completions/mean_terminated_length": 365.375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.5084, "grad_norm": 0.07548683881759644, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29831176.0, "reward": 0.6553915143013, "reward_std": 0.24372483789920807, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24828124046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 2542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 361.375, "completions/mean_terminated_length": 361.375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.5086, "grad_norm": 0.00027647471870295703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29840958.0, "reward": 1.499257206916809, "reward_std": 0.000594696553889662, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985125064849854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 2543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 387.0, "completions/max_terminated_length": 387.0, "completions/mean_length": 294.9375, "completions/mean_terminated_length": 294.9375, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.5088, "grad_norm": 0.1735243946313858, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29849581.0, "reward": 0.7711732983589172, "reward_std": 0.4347483515739441, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29234373569488525, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 2544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 205.5, "completions/mean_terminated_length": 205.5, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 0.509, "grad_norm": 0.35159584879875183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29856373.0, "reward": 1.2004543542861938, "reward_std": 0.4578687250614166, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7134062051773071, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 2545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 284.0, "completions/max_terminated_length": 284.0, "completions/mean_length": 241.0625, "completions/mean_terminated_length": 241.0625, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.5092, "grad_norm": 0.10994409769773483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29863830.0, "reward": 1.4379419088363647, "reward_std": 0.24746458232402802, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9383811950683594, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9881250262260437, "step": 2546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 301.625, "completions/mean_terminated_length": 301.625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.5094, "grad_norm": 0.2024906724691391, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29872928.0, "reward": 1.0190479755401611, "reward_std": 0.49591711163520813, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5380937457084656, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 2547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 425.3125, "completions/mean_terminated_length": 425.3125, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.5096, "grad_norm": 0.13357307016849518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29884629.0, "reward": 0.9284133911132812, "reward_std": 0.463412880897522, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4818250238895416, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9225000143051147, "step": 2548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 463.875, "completions/mean_terminated_length": 463.875, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.5098, "grad_norm": 0.05323093757033348, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29896403.0, "reward": 1.4374290704727173, "reward_std": 0.24732810258865356, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9373562335968018, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 2549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 589.4375, "completions/mean_terminated_length": 560.4666748046875, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.51, "grad_norm": 0.08863192796707153, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29909978.0, "reward": 1.2115508317947388, "reward_std": 0.4811156690120697, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7981000542640686, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8237500190734863, "step": 2550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 412.125, "completions/mean_terminated_length": 412.125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.5102, "grad_norm": 0.09298282861709595, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29920316.0, "reward": 0.665075957775116, "reward_std": 0.3255765438079834, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20514999330043793, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 740.3125, "completions/mean_terminated_length": 721.4000244140625, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.5104, "grad_norm": 0.10391367226839066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29936929.0, "reward": 0.8513163328170776, "reward_std": 0.4351818561553955, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5151312351226807, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6268749833106995, "step": 2552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 434.0, "completions/mean_terminated_length": 434.0, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5106, "grad_norm": 0.21186937391757965, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29948113.0, "reward": 1.320766568183899, "reward_std": 0.38302719593048096, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290312886238098, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 906.9375, "completions/mean_terminated_length": 711.8333740234375, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.5108, "grad_norm": 0.0932321697473526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29967136.0, "reward": 0.21562542021274567, "reward_std": 0.2921905219554901, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05624999850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.34375, "step": 2554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 664.0, "completions/max_terminated_length": 664.0, "completions/mean_length": 457.4375, "completions/mean_terminated_length": 457.4375, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.511, "grad_norm": 0.11298218369483948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29978631.0, "reward": 1.287775993347168, "reward_std": 0.3915720283985138, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8255499601364136, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 2555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 527.75, "completions/mean_terminated_length": 527.75, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.5112, "grad_norm": 0.12141653150320053, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29992291.0, "reward": 0.8201446533203125, "reward_std": 0.41994044184684753, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39028748869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8862500190734863, "step": 2556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 259.5, "completions/mean_terminated_length": 259.5, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.5114, "grad_norm": 0.2793552279472351, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29999931.0, "reward": 1.0241665840148926, "reward_std": 0.49100732803344727, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5483312606811523, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 510.0, "completions/mean_terminated_length": 510.0, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.5116, "grad_norm": 0.029311079531908035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30012395.0, "reward": 0.8503910899162292, "reward_std": 0.10277049988508224, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7007812261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 2558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 497.5, "completions/mean_terminated_length": 497.5, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.5118, "grad_norm": 0.10250900685787201, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30024035.0, "reward": 1.225935459136963, "reward_std": 0.43075528740882874, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7643687725067139, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9118750095367432, "step": 2559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 259.875, "completions/mean_terminated_length": 259.875, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.512, "grad_norm": 0.1762237250804901, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30031665.0, "reward": 1.0753417015075684, "reward_std": 0.4974096119403839, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5881812572479248, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 2560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 725.6875, "completions/mean_terminated_length": 590.0909423828125, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.5122, "grad_norm": 0.11120189726352692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30047196.0, "reward": 0.6115633249282837, "reward_std": 0.5807698369026184, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28562498092651367, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.65625, "step": 2561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 423.875, "completions/mean_terminated_length": 423.875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.5124, "grad_norm": 0.10287731140851974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30057674.0, "reward": 0.9128851890563965, "reward_std": 0.4313034415245056, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.513268768787384, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8456250429153442, "step": 2562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 776.125, "completions/mean_terminated_length": 693.5, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.5126, "grad_norm": 0.05334554612636566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30073788.0, "reward": 0.48070403933525085, "reward_std": 0.37314504384994507, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14890623092651367, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.703125, "step": 2563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 824.875, "completions/mean_terminated_length": 705.4000244140625, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.5128, "grad_norm": 0.05669035017490387, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30091050.0, "reward": 0.34054434299468994, "reward_std": 0.2726011872291565, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05608749762177467, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5962499976158142, "step": 2564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 713.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 558.3125, "completions/mean_terminated_length": 558.3125, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.513, "grad_norm": 0.08430760353803635, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30104959.0, "reward": 0.8691228628158569, "reward_std": 0.45323288440704346, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4257437288761139, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 2565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 317.0, "completions/max_terminated_length": 317.0, "completions/mean_length": 256.0, "completions/mean_terminated_length": 256.0, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.5132, "grad_norm": 0.21967177093029022, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30113007.0, "reward": 1.0177323818206787, "reward_std": 0.49718791246414185, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.535462498664856, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 497.9375, "completions/mean_terminated_length": 497.9375, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.5134, "grad_norm": 0.0861244723200798, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30125790.0, "reward": 0.6642979383468628, "reward_std": 0.32642242312431335, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20359374582767487, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 815.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 471.125, "completions/mean_terminated_length": 471.125, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.5136, "grad_norm": 0.11971012502908707, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30137360.0, "reward": 0.719022810459137, "reward_std": 0.3870950937271118, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2505437433719635, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 530.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 384.1875, "completions/mean_terminated_length": 384.1875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.5138, "grad_norm": 0.14656178653240204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30148579.0, "reward": 0.5978761315345764, "reward_std": 0.24129679799079895, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13324999809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 2569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 408.25, "completions/mean_terminated_length": 408.25, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.514, "grad_norm": 0.10503584891557693, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30160039.0, "reward": 0.8317195177078247, "reward_std": 0.36100950837135315, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.47593751549720764, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.793749988079071, "step": 2570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 739.5, "completions/mean_terminated_length": 673.84619140625, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.5142, "grad_norm": 0.1076873317360878, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30175951.0, "reward": 0.7959349751472473, "reward_std": 0.5261558294296265, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5293687582015991, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5743750333786011, "step": 2571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 697.625, "completions/mean_terminated_length": 697.625, "completions/min_length": 560.0, "completions/min_terminated_length": 560.0, "epoch": 0.5144, "grad_norm": 0.07951551675796509, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30190905.0, "reward": 1.074951171875, "reward_std": 0.4970804452896118, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.587399959564209, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 363.375, "completions/mean_terminated_length": 363.375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.5146, "grad_norm": 0.1403084248304367, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30201199.0, "reward": 0.8114416003227234, "reward_std": 0.4331375062465668, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3103812336921692, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9106249809265137, "step": 2573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 388.75, "completions/mean_terminated_length": 388.75, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.5148, "grad_norm": 0.22862260043621063, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30211467.0, "reward": 1.0178792476654053, "reward_std": 0.49733978509902954, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5357562303543091, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 573.9375, "completions/mean_terminated_length": 543.933349609375, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.515, "grad_norm": 0.10441488772630692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30224730.0, "reward": 0.6955478191375732, "reward_std": 0.42177054286003113, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26609376072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 406.375, "completions/mean_terminated_length": 406.375, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.5152, "grad_norm": 0.1684027910232544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30235568.0, "reward": 1.2108228206634521, "reward_std": 0.4433060586452484, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7341437339782715, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 2576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 982.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 599.75, "completions/mean_terminated_length": 599.75, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.5154, "grad_norm": 0.022610392421483994, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30249132.0, "reward": 0.5708572268486023, "reward_std": 0.08105108141899109, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.141712486743927, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 2577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 408.9375, "completions/mean_terminated_length": 408.9375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.5156, "grad_norm": 0.018588513135910034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30260251.0, "reward": 0.5503138303756714, "reward_std": 0.07743172347545624, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10062499344348907, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 578.375, "completions/mean_terminated_length": 578.375, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.5158, "grad_norm": 0.06092444062232971, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30274145.0, "reward": 0.6716322898864746, "reward_std": 0.27657607197761536, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2807624936103821, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8325000405311584, "step": 2579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 642.3125, "completions/mean_terminated_length": 616.86669921875, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.516, "grad_norm": 0.11926218867301941, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30288310.0, "reward": 0.9038947224617004, "reward_std": 0.5131503939628601, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4952875077724457, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8237500190734863, "step": 2580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 708.0, "completions/max_terminated_length": 708.0, "completions/mean_length": 489.75, "completions/mean_terminated_length": 489.75, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.5162, "grad_norm": 0.170576810836792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30299826.0, "reward": 0.902019739151001, "reward_std": 0.4776742458343506, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42903751134872437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 709.625, "completions/mean_terminated_length": 664.7142944335938, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.5164, "grad_norm": 0.032602861523628235, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30315532.0, "reward": 0.4801570773124695, "reward_std": 0.187434121966362, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08531250059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 2582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 316.0, "completions/max_terminated_length": 316.0, "completions/mean_length": 250.9375, "completions/mean_terminated_length": 250.9375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.5166, "grad_norm": 0.09232039004564285, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30323435.0, "reward": 0.6821105480194092, "reward_std": 0.22254984080791473, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.36421874165534973, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6781250238418579, "step": 2583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 257.25, "completions/mean_terminated_length": 257.25, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.5168, "grad_norm": 0.09632369875907898, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30331119.0, "reward": 0.6051573157310486, "reward_std": 0.23893482983112335, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14781250059604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 372.625, "completions/mean_terminated_length": 372.625, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.517, "grad_norm": 0.07180187851190567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30340617.0, "reward": 0.5838573575019836, "reward_std": 0.24473661184310913, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1052125096321106, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 381.4375, "completions/mean_terminated_length": 381.4375, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.5172, "grad_norm": 0.0066210003569722176, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30351584.0, "reward": 0.5384229421615601, "reward_std": 0.01895567961037159, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0768437460064888, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 2586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 366.1875, "completions/mean_terminated_length": 366.1875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.5174, "grad_norm": 0.00849505327641964, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30361379.0, "reward": 0.5406917929649353, "reward_std": 0.024586334824562073, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08138124644756317, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 255.8125, "completions/mean_terminated_length": 255.8125, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.5176, "grad_norm": 0.00037374242674559355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30369328.0, "reward": 1.4992635250091553, "reward_std": 0.0006027987692505121, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985250234603882, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 2588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 248.3125, "completions/mean_terminated_length": 248.3125, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.5178, "grad_norm": 0.23241335153579712, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30377237.0, "reward": 1.3184354305267334, "reward_std": 0.3883736729621887, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8243687152862549, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 293.0625, "completions/mean_terminated_length": 293.0625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.518, "grad_norm": 0.09150054305791855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30385894.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 279.25, "completions/mean_terminated_length": 279.25, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.5182, "grad_norm": 0.17745018005371094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30394762.0, "reward": 1.1326011419296265, "reward_std": 0.4892868995666504, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6402000188827515, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9749999642372131, "step": 2591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 583.3125, "completions/mean_terminated_length": 553.933349609375, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.5184, "grad_norm": 0.054919373244047165, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30408751.0, "reward": 0.6843756437301636, "reward_std": 0.24782800674438477, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4312500059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6437499523162842, "step": 2592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 616.75, "completions/mean_terminated_length": 616.75, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.5186, "grad_norm": 0.1134057268500328, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30422859.0, "reward": 1.1624010801315308, "reward_std": 0.46306082606315613, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6998000144958496, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 2593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 473.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 331.3125, "completions/mean_terminated_length": 331.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.5188, "grad_norm": 0.07276555895805359, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30431808.0, "reward": 0.605079174041748, "reward_std": 0.2386225014925003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 321.3125, "completions/mean_terminated_length": 321.3125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.519, "grad_norm": 0.23996324837207794, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30441349.0, "reward": 0.9043633341789246, "reward_std": 0.4390557110309601, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4962249994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 2595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 218.9375, "completions/mean_terminated_length": 218.9375, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 0.5192, "grad_norm": 0.22345741093158722, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30449044.0, "reward": 0.779666543006897, "reward_std": 0.42918357253074646, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3093312382698059, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 2596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 431.5625, "completions/mean_terminated_length": 431.5625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.5194, "grad_norm": 0.09254846721887589, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30459677.0, "reward": 1.212500810623169, "reward_std": 0.30597206950187683, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.9249999523162842, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5687500238418579, "step": 2597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 654.625, "completions/mean_terminated_length": 531.5, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.5196, "grad_norm": 0.10760330408811569, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30475159.0, "reward": 0.785938024520874, "reward_std": 0.5429363250732422, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6343749761581421, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.32499998807907104, "step": 2598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 628.6875, "completions/mean_terminated_length": 602.3333740234375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.5198, "grad_norm": 0.07480312883853912, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30490594.0, "reward": 0.6841260194778442, "reward_std": 0.4261188805103302, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2432500123977661, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 2599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 562.9375, "completions/mean_terminated_length": 562.9375, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.52, "grad_norm": 0.050043489784002304, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30503809.0, "reward": 0.6498446464538574, "reward_std": 0.25477394461631775, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23718750476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 2600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 473.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 315.5625, "completions/mean_terminated_length": 315.5625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.5202, "grad_norm": 0.011137680150568485, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30512442.0, "reward": 0.5346043109893799, "reward_std": 0.027632994577288628, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06920625269412994, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 424.8125, "completions/mean_terminated_length": 424.8125, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.5204, "grad_norm": 0.00020561242126859725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30523223.0, "reward": 1.4992823600769043, "reward_std": 0.0006225122488103807, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985625147819519, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 407.4375, "completions/mean_terminated_length": 407.4375, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.5206, "grad_norm": 0.16545777022838593, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30533726.0, "reward": 0.7873353958129883, "reward_std": 0.42514219880104065, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3246687352657318, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 279.25, "completions/mean_terminated_length": 279.25, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.5208, "grad_norm": 0.24710002541542053, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30541730.0, "reward": 1.2509762048721313, "reward_std": 0.44483593106269836, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7519500255584717, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.98375004529953, "step": 2604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 386.8125, "completions/mean_terminated_length": 386.8125, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.521, "grad_norm": 0.13504891097545624, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30551711.0, "reward": 1.2568355798721313, "reward_std": 0.4339340925216675, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7636687755584717, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 2605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 652.5625, "completions/mean_terminated_length": 627.800048828125, "completions/min_length": 460.0, "completions/min_terminated_length": 460.0, "epoch": 0.5212, "grad_norm": 0.07681235671043396, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30566072.0, "reward": 1.3175759315490723, "reward_std": 0.43101298809051514, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8851499557495117, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8337500095367432, "step": 2606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 361.5625, "completions/mean_terminated_length": 361.5625, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.5214, "grad_norm": 0.00020231753296684474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30575713.0, "reward": 1.4997355937957764, "reward_std": 0.0004889005795121193, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9994688034057617, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 2607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 550.625, "completions/mean_terminated_length": 550.625, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.5216, "grad_norm": 0.12605711817741394, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30589067.0, "reward": 0.9615479707717896, "reward_std": 0.4899054765701294, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4855937659740448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 2608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 296.5, "completions/mean_terminated_length": 296.5, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.5218, "grad_norm": 0.3917796313762665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30597523.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 790.5, "completions/mean_terminated_length": 757.1428833007812, "completions/min_length": 614.0, "completions/min_terminated_length": 614.0, "epoch": 0.522, "grad_norm": 0.07466527819633484, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30614283.0, "reward": 0.5813289284706116, "reward_std": 0.2944495379924774, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2876562476158142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.640625, "step": 2610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 491.25, "completions/mean_terminated_length": 491.25, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.5222, "grad_norm": 0.11149778962135315, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30626127.0, "reward": 1.2019414901733398, "reward_std": 0.45482364296913147, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.716381311416626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 331.625, "completions/mean_terminated_length": 331.625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.5224, "grad_norm": 0.2218063920736313, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30635417.0, "reward": 1.3206915855407715, "reward_std": 0.38298988342285156, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8288812637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 381.4375, "completions/mean_terminated_length": 381.4375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.5226, "grad_norm": 0.17536038160324097, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30645328.0, "reward": 1.0282825231552124, "reward_std": 0.48878875374794006, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5565624833106995, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 492.125, "completions/mean_terminated_length": 492.125, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.5228, "grad_norm": 0.19876867532730103, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30657506.0, "reward": 1.2064073085784912, "reward_std": 0.5006471276283264, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7253124713897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 2614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 845.625, "completions/mean_terminated_length": 804.4615478515625, "completions/min_length": 585.0, "completions/min_terminated_length": 585.0, "epoch": 0.523, "grad_norm": 0.12113998085260391, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30675116.0, "reward": 0.9695727229118347, "reward_std": 0.6067436933517456, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6266437768936157, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7181249856948853, "step": 2615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 676.75, "completions/mean_terminated_length": 627.1428833007812, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.5232, "grad_norm": 0.1301417201757431, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30689624.0, "reward": 0.9036601781845093, "reward_std": 0.5954249501228333, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5573187470436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7800000309944153, "step": 2616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 256.5, "completions/mean_terminated_length": 256.5, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.5234, "grad_norm": 0.3756106197834015, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30697728.0, "reward": 1.0208479166030884, "reward_std": 0.49401360750198364, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5416937470436096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 2617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 646.75, "completions/mean_terminated_length": 646.75, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.5236, "grad_norm": 0.10953379422426224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30711772.0, "reward": 0.7723199129104614, "reward_std": 0.43379026651382446, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2946375012397766, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 2618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 594.6875, "completions/mean_terminated_length": 594.6875, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.5238, "grad_norm": 0.11687829345464706, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30725751.0, "reward": 1.3178730010986328, "reward_std": 0.38978931307792664, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8232437372207642, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 2619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 446.5, "completions/mean_terminated_length": 408.0000305175781, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.524, "grad_norm": 0.20919229090213776, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30736591.0, "reward": 1.1602352857589722, "reward_std": 0.49261656403541565, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6954687833786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.765625, "step": 2620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 618.625, "completions/mean_terminated_length": 591.6000366210938, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.5242, "grad_norm": 0.12737400829792023, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30750313.0, "reward": 0.7152228355407715, "reward_std": 0.42028459906578064, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.30544376373291016, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 2621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 604.0, "completions/max_terminated_length": 604.0, "completions/mean_length": 436.0, "completions/mean_terminated_length": 436.0, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.5244, "grad_norm": 0.12865866720676422, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30761417.0, "reward": 0.8887386322021484, "reward_std": 0.4074413776397705, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5274749994277954, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7549999952316284, "step": 2622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 382.5, "completions/mean_terminated_length": 382.5, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.5246, "grad_norm": 0.12710042297840118, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30771697.0, "reward": 0.7220197916030884, "reward_std": 0.3856043815612793, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2565374970436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 2623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 707.9375, "completions/mean_terminated_length": 686.86669921875, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.5248, "grad_norm": 0.04502549394965172, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30786896.0, "reward": 0.5449227690696716, "reward_std": 0.18448267877101898, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15234375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 2624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 410.0, "completions/max_terminated_length": 410.0, "completions/mean_length": 331.625, "completions/mean_terminated_length": 331.625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.525, "grad_norm": 0.2345341145992279, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30795946.0, "reward": 0.9508635401725769, "reward_std": 0.5001057386398315, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46422499418258667, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 2625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 437.8125, "completions/mean_terminated_length": 437.8125, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.5252, "grad_norm": 0.10077914595603943, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30806583.0, "reward": 1.3208447694778442, "reward_std": 0.3830660879611969, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291875123977661, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 2626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 563.9375, "completions/mean_terminated_length": 533.2667236328125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.5254, "grad_norm": 0.14138221740722656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30819382.0, "reward": 0.9496445655822754, "reward_std": 0.49053430557250977, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5867875218391418, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7549999952316284, "step": 2627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 374.5, "completions/mean_terminated_length": 374.5, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.5256, "grad_norm": 0.18469814956188202, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30829614.0, "reward": 1.01616370677948, "reward_std": 0.49928006529808044, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5323249697685242, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 700.0625, "completions/mean_terminated_length": 678.4666748046875, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.5258, "grad_norm": 0.10153774172067642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30846079.0, "reward": 0.8347541093826294, "reward_std": 0.49759969115257263, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4195062518119812, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 2629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 497.75, "completions/mean_terminated_length": 462.66668701171875, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.526, "grad_norm": 0.15828411281108856, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30858523.0, "reward": 1.1677196025848389, "reward_std": 0.5230008959770203, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.710437536239624, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8949999809265137, "step": 2630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 530.5, "completions/mean_terminated_length": 530.5, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.5262, "grad_norm": 0.1187969446182251, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30870723.0, "reward": 1.1374760866165161, "reward_std": 0.4828333556652069, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6499499082565308, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9711875319480896, "step": 2631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 834.125, "completions/mean_terminated_length": 720.2000122070312, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.5264, "grad_norm": 0.08551058173179626, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30887989.0, "reward": 0.4378131330013275, "reward_std": 0.4193057119846344, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18812498450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5499999523162842, "step": 2632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 477.0, "completions/mean_terminated_length": 398.8571472167969, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.5266, "grad_norm": 0.15309341251850128, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30899477.0, "reward": 1.2206132411956787, "reward_std": 0.5422794818878174, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.816224992275238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 2633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 324.5625, "completions/mean_terminated_length": 324.5625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.5268, "grad_norm": 0.00028662270051427186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30908734.0, "reward": 1.4995794296264648, "reward_std": 0.0005786879337392747, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991562366485596, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 2634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 316.1875, "completions/mean_terminated_length": 316.1875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.527, "grad_norm": 0.14897775650024414, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30917201.0, "reward": 0.5752259492874146, "reward_std": 0.28323665261268616, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1504499912261963, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 2635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 409.375, "completions/mean_terminated_length": 409.375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.5272, "grad_norm": 0.1299523115158081, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30927623.0, "reward": 1.319044828414917, "reward_std": 0.387008935213089, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8255875110626221, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 2636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 534.9375, "completions/mean_terminated_length": 534.9375, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.5274, "grad_norm": 0.06017230078577995, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30939910.0, "reward": 0.817188024520874, "reward_std": 0.153232604265213, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6343749761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.550000011920929, "step": 2637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 218.625, "completions/mean_terminated_length": 218.625, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.5276, "grad_norm": 0.07324951887130737, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30946992.0, "reward": 0.9412506818771362, "reward_std": 0.15469828248023987, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8825000524520874, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.16249999403953552, "step": 2638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 330.375, "completions/mean_terminated_length": 330.375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.5278, "grad_norm": 0.00040673339390195906, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30956582.0, "reward": 1.498582124710083, "reward_std": 0.0009902730816975236, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9971624612808228, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 801.5625, "completions/mean_terminated_length": 700.45458984375, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.528, "grad_norm": 0.10291128605604172, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30974063.0, "reward": 0.41984429955482483, "reward_std": 0.40933114290237427, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2771874964237213, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.375, "step": 2640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 380.5625, "completions/mean_terminated_length": 380.5625, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.5282, "grad_norm": 0.1481332778930664, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30984888.0, "reward": 0.7832822799682617, "reward_std": 0.4269797205924988, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3165625035762787, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 2641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 407.4375, "completions/mean_terminated_length": 407.4375, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.5284, "grad_norm": 0.00020641597802750766, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30995215.0, "reward": 1.4997355937957764, "reward_std": 0.0004889005795121193, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9994688034057617, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 2642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 417.4375, "completions/mean_terminated_length": 417.4375, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.5286, "grad_norm": 0.08698611706495285, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31005878.0, "reward": 0.6619884371757507, "reward_std": 0.32748112082481384, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19897499680519104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 2643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 408.1875, "completions/mean_terminated_length": 408.1875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.5288, "grad_norm": 0.0877670869231224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31017289.0, "reward": 1.378257155418396, "reward_std": 0.3294990360736847, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8815125226974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 455.9375, "completions/mean_terminated_length": 455.9375, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.529, "grad_norm": 0.08302636444568634, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31028408.0, "reward": 0.663960337638855, "reward_std": 0.32625994086265564, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20291873812675476, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 324.125, "completions/mean_terminated_length": 324.125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.5292, "grad_norm": 0.2338494509458542, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31037818.0, "reward": 1.0804450511932373, "reward_std": 0.49139901995658875, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5983874797821045, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 2646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 340.625, "completions/mean_terminated_length": 340.625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.5294, "grad_norm": 0.11362726241350174, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31049780.0, "reward": 0.5971823334693909, "reward_std": 0.24121001362800598, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13186249136924744, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 632.0, "completions/mean_terminated_length": 632.0, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.5296, "grad_norm": 0.03262671083211899, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31064724.0, "reward": 0.9057034254074097, "reward_std": 0.1352117359638214, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8114062547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3343749940395355, "step": 2648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 871.0, "completions/mean_terminated_length": 820.0, "completions/min_length": 542.0, "completions/min_terminated_length": 542.0, "epoch": 0.5298, "grad_norm": 0.09242156147956848, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31083332.0, "reward": 0.5754692554473877, "reward_std": 0.3918849527835846, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4009374976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.41874998807907104, "step": 2649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 262.0, "completions/max_terminated_length": 262.0, "completions/mean_length": 237.625, "completions/mean_terminated_length": 237.625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.53, "grad_norm": 0.2541945278644562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31090670.0, "reward": 1.0725228786468506, "reward_std": 0.499815970659256, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5825434923171997, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.968000054359436, "step": 2650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 495.0, "completions/mean_terminated_length": 495.0, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.5302, "grad_norm": 0.12105485796928406, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31102846.0, "reward": 0.9101571440696716, "reward_std": 0.46442049741744995, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4453125, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 2651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 815.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 546.5625, "completions/mean_terminated_length": 546.5625, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.5304, "grad_norm": 0.00013724446762353182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31115175.0, "reward": 1.499038577079773, "reward_std": 0.0005171343218535185, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998075008392334, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 2652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 369.0, "completions/mean_terminated_length": 369.0, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.5306, "grad_norm": 0.12862592935562134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31125527.0, "reward": 0.7245165705680847, "reward_std": 0.38442206382751465, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2615312337875366, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 428.375, "completions/mean_terminated_length": 428.375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5308, "grad_norm": 0.08640559017658234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31137149.0, "reward": 0.6862510442733765, "reward_std": 0.31822606921195984, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2475000023841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 2654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 386.75, "completions/mean_terminated_length": 386.75, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.531, "grad_norm": 0.17709453403949738, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31146905.0, "reward": 0.8421509265899658, "reward_std": 0.45761796832084656, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3718000054359436, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 464.75, "completions/mean_terminated_length": 464.75, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.5312, "grad_norm": 0.1437147855758667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31158565.0, "reward": 1.1689603328704834, "reward_std": 0.45290955901145935, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.712918758392334, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 2656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 406.125, "completions/mean_terminated_length": 406.125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.5314, "grad_norm": 0.22536683082580566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31169127.0, "reward": 0.7801449298858643, "reward_std": 0.4294920861721039, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3102875053882599, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 2657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 273.0, "completions/mean_terminated_length": 273.0, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.5316, "grad_norm": 0.00017109733016695827, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31176807.0, "reward": 1.4989478588104248, "reward_std": 0.00042326669790782034, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.997893750667572, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 406.9375, "completions/mean_terminated_length": 406.9375, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.5318, "grad_norm": 0.012574944645166397, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31187702.0, "reward": 0.5507509708404541, "reward_std": 0.05035773292183876, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1014999970793724, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 2659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 441.25, "completions/mean_terminated_length": 441.25, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.532, "grad_norm": 0.14200517535209656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31198730.0, "reward": 0.7455477714538574, "reward_std": 0.3705313503742218, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3035937547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.909375011920929, "step": 2660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 691.5, "completions/mean_terminated_length": 644.0, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.5322, "grad_norm": 0.09589170664548874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31214162.0, "reward": 0.6196103096008301, "reward_std": 0.39878466725349426, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23921874165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 2661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 454.0, "completions/mean_terminated_length": 416.0000305175781, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.5324, "grad_norm": 0.1420820951461792, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31225442.0, "reward": 0.9252946376800537, "reward_std": 0.5388302803039551, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47558748722076416, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8975000381469727, "step": 2662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 708.8125, "completions/mean_terminated_length": 636.0769653320312, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.5326, "grad_norm": 0.058130647987127304, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31241567.0, "reward": 0.4458601474761963, "reward_std": 0.22120855748653412, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 2663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 882.9375, "completions/mean_terminated_length": 773.2222290039062, "completions/min_length": 505.0, "completions/min_terminated_length": 505.0, "epoch": 0.5328, "grad_norm": 0.0695873275399208, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31260558.0, "reward": 0.38835978507995605, "reward_std": 0.3771863281726837, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21421873569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.40937498211860657, "step": 2664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 262.0, "completions/max_terminated_length": 262.0, "completions/mean_length": 226.5625, "completions/mean_terminated_length": 226.5625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.533, "grad_norm": 0.26708266139030457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31267831.0, "reward": 1.129563570022583, "reward_std": 0.4932810962200165, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6341249942779541, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "step": 2665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 425.625, "completions/mean_terminated_length": 425.625, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.5332, "grad_norm": 0.09650718420743942, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31278753.0, "reward": 1.4398261308670044, "reward_std": 0.23762080073356628, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421499967575073, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 2666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 660.25, "completions/mean_terminated_length": 660.25, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.5334, "grad_norm": 0.03057621233165264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31293381.0, "reward": 0.5749228000640869, "reward_std": 0.08155056834220886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1498437523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 2667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 418.75, "completions/mean_terminated_length": 418.75, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.5336, "grad_norm": 0.16635769605636597, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31304065.0, "reward": 0.7936729192733765, "reward_std": 0.42316296696662903, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3373437523841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 472.625, "completions/mean_terminated_length": 472.625, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.5338, "grad_norm": 0.0756409540772438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31315675.0, "reward": 0.6625633239746094, "reward_std": 0.27055585384368896, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2626250088214874, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8449999690055847, "step": 2669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 808.6875, "completions/mean_terminated_length": 759.0, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.534, "grad_norm": 0.08552682399749756, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31332758.0, "reward": 0.6505223512649536, "reward_std": 0.417804479598999, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.42604371905326843, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5131250023841858, "step": 2670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 663.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 526.125, "completions/mean_terminated_length": 526.125, "completions/min_length": 436.0, "completions/min_terminated_length": 436.0, "epoch": 0.5342, "grad_norm": 0.10944665968418121, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31345208.0, "reward": 1.0142196416854858, "reward_std": 0.5008174180984497, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5284374952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 2671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 788.875, "completions/mean_terminated_length": 710.5, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.5344, "grad_norm": 0.056293390691280365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31362502.0, "reward": 0.5232037305831909, "reward_std": 0.3586268424987793, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.29640623927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.515625, "step": 2672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 986.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 478.1875, "completions/mean_terminated_length": 478.1875, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.5346, "grad_norm": 0.16161052882671356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31374665.0, "reward": 1.0799227952957153, "reward_std": 0.4710480272769928, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5973438024520874, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 2673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 362.5, "completions/mean_terminated_length": 362.5, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5348, "grad_norm": 0.1402936428785324, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31385025.0, "reward": 1.201951026916504, "reward_std": 0.45483019948005676, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7164000272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 355.9375, "completions/mean_terminated_length": 355.9375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.535, "grad_norm": 0.05598319694399834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31395040.0, "reward": 1.4600634574890137, "reward_std": 0.15601769089698792, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9826249480247498, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9324374794960022, "step": 2675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 601.0625, "completions/mean_terminated_length": 540.6428833007812, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.5352, "grad_norm": 0.136015385389328, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31408577.0, "reward": 0.8630476593971252, "reward_std": 0.5245630145072937, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5385937690734863, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.671875, "step": 2676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 654.5625, "completions/mean_terminated_length": 629.933349609375, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.5354, "grad_norm": 0.10335367172956467, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31423194.0, "reward": 0.769828736782074, "reward_std": 0.38236841559410095, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.41465625166893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6231250166893005, "step": 2677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 844.1875, "completions/mean_terminated_length": 818.5000610351562, "completions/min_length": 602.0, "completions/min_terminated_length": 602.0, "epoch": 0.5356, "grad_norm": 0.07265018671751022, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31440861.0, "reward": 0.824988067150116, "reward_std": 0.4460104703903198, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.587475061416626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5362499952316284, "step": 2678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 660.625, "completions/mean_terminated_length": 636.4000244140625, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.5358, "grad_norm": 0.09316230565309525, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31455175.0, "reward": 0.8661603331565857, "reward_std": 0.45231059193611145, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4198187589645386, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 2679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 451.1875, "completions/mean_terminated_length": 451.1875, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.536, "grad_norm": 0.14112891256809235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31466074.0, "reward": 0.6166415214538574, "reward_std": 0.2365012764930725, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1707812398672104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "step": 2680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 704.8125, "completions/mean_terminated_length": 659.2142944335938, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.5362, "grad_norm": 0.06925100833177567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31481271.0, "reward": 0.5819538831710815, "reward_std": 0.27690255641937256, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2889062464237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6781249642372131, "step": 2681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 657.8125, "completions/mean_terminated_length": 657.8125, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.5364, "grad_norm": 0.10417898744344711, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31495956.0, "reward": 0.6887383460998535, "reward_std": 0.32596123218536377, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2524750232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 2682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 509.0, "completions/max_terminated_length": 509.0, "completions/mean_length": 383.125, "completions/mean_terminated_length": 383.125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5366, "grad_norm": 0.13519977033138275, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31505990.0, "reward": 1.2002758979797363, "reward_std": 0.4577445089817047, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7130500078201294, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 455.5, "completions/mean_terminated_length": 455.5, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.5368, "grad_norm": 0.14369043707847595, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31517774.0, "reward": 1.0778510570526123, "reward_std": 0.4935058057308197, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5931999683380127, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 332.375, "completions/mean_terminated_length": 332.375, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.537, "grad_norm": 0.2050386369228363, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31526868.0, "reward": 1.2534229755401611, "reward_std": 0.43492695689201355, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7568437457084656, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 2685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 655.125, "completions/mean_terminated_length": 487.4545593261719, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.5372, "grad_norm": 0.4148937165737152, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31541158.0, "reward": 0.742422342300415, "reward_std": 0.5720281004905701, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6098437309265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.32187503576278687, "step": 2686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 929.8125, "completions/mean_terminated_length": 873.2999877929688, "completions/min_length": 693.0, "completions/min_terminated_length": 693.0, "epoch": 0.5374, "grad_norm": 0.08027570694684982, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31560435.0, "reward": 0.3942067623138428, "reward_std": 0.3268106281757355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16341249644756317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5299999713897705, "step": 2687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 903.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 663.1875, "completions/mean_terminated_length": 663.1875, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.5376, "grad_norm": 0.09875673055648804, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31575446.0, "reward": 0.7139853835105896, "reward_std": 0.32637736201286316, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30296874046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 2688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 427.6875, "completions/mean_terminated_length": 427.6875, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.5378, "grad_norm": 0.11773227900266647, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31586625.0, "reward": 0.9032571315765381, "reward_std": 0.47658780217170715, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4315125048160553, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 393.375, "completions/mean_terminated_length": 393.375, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.538, "grad_norm": 0.26596105098724365, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31596871.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 936.125, "completions/mean_terminated_length": 848.25, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.5382, "grad_norm": 0.09575585275888443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31617161.0, "reward": 0.3253909945487976, "reward_std": 0.3916497230529785, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21328124403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.36250001192092896, "step": 2691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 562.3125, "completions/mean_terminated_length": 562.3125, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.5384, "grad_norm": 0.06095670908689499, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31629710.0, "reward": 0.6620291471481323, "reward_std": 0.3269628584384918, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1990562528371811, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249380111694, "step": 2692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 438.6875, "completions/mean_terminated_length": 438.6875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.5386, "grad_norm": 0.06213819980621338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31640889.0, "reward": 0.714453935623169, "reward_std": 0.19955644011497498, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.42890626192092896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6656249761581421, "step": 2693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 474.9375, "completions/mean_terminated_length": 474.9375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5388, "grad_norm": 0.08031316101551056, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31652232.0, "reward": 0.5911728143692017, "reward_std": 0.21696248650550842, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2448437511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 2694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 476.5, "completions/mean_terminated_length": 398.2857360839844, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.539, "grad_norm": 0.1520436406135559, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31663872.0, "reward": 0.81937575340271, "reward_std": 0.5279877185821533, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45125001668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 2695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 415.1875, "completions/mean_terminated_length": 415.1875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.5392, "grad_norm": 0.08286735415458679, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31674371.0, "reward": 0.7861727476119995, "reward_std": 0.28005000948905945, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5098437666893005, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6281250715255737, "step": 2696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 652.6875, "completions/mean_terminated_length": 599.6428833007812, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.5394, "grad_norm": 0.08330171555280685, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31688350.0, "reward": 0.7851572036743164, "reward_std": 0.5319567322731018, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3828125, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.824999988079071, "step": 2697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 461.375, "completions/mean_terminated_length": 461.375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.5396, "grad_norm": 0.01029143389314413, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31699684.0, "reward": 0.5656259059906006, "reward_std": 0.03567207232117653, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13124999403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 2698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 659.9375, "completions/mean_terminated_length": 607.9285888671875, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.5398, "grad_norm": 0.09171433746814728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31715699.0, "reward": 0.6900789737701416, "reward_std": 0.4603666663169861, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3176562488079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 2699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 523.3125, "completions/mean_terminated_length": 523.3125, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.54, "grad_norm": 0.06295520067214966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31728760.0, "reward": 0.6360164880752563, "reward_std": 0.2555769681930542, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2095312476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 2700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 361.5625, "completions/mean_terminated_length": 361.5625, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.5402, "grad_norm": 0.11107608675956726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31738641.0, "reward": 1.4365605115890503, "reward_std": 0.24974939227104187, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9356187582015991, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 722.4375, "completions/mean_terminated_length": 652.84619140625, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.5404, "grad_norm": 0.13148394227027893, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31754264.0, "reward": 0.7026538848876953, "reward_std": 0.4916839897632599, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.40530624985694885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6681250333786011, "step": 2702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 548.1875, "completions/mean_terminated_length": 548.1875, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.5406, "grad_norm": 0.10263467580080032, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31767259.0, "reward": 1.2187789678573608, "reward_std": 0.43639469146728516, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.750056266784668, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9293749928474426, "step": 2703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 472.9375, "completions/mean_terminated_length": 472.9375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.5408, "grad_norm": 0.05662725120782852, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31779114.0, "reward": 0.5911386013031006, "reward_std": 0.2428576946258545, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11977499723434448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 309.5, "completions/mean_terminated_length": 309.5, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.541, "grad_norm": 0.10879837721586227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31787970.0, "reward": 0.6328135132789612, "reward_std": 0.2576128840446472, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.203125, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 2705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 642.0625, "completions/mean_terminated_length": 587.5, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.5412, "grad_norm": 0.10186313092708588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31802803.0, "reward": 0.609382152557373, "reward_std": 0.35862332582473755, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.28126251697540283, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6999937295913696, "step": 2706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 472.125, "completions/mean_terminated_length": 435.3333435058594, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.5414, "grad_norm": 0.13395783305168152, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31814325.0, "reward": 1.195847749710083, "reward_std": 0.49896571040153503, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7041937708854675, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8331249952316284, "step": 2707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 362.6875, "completions/mean_terminated_length": 362.6875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.5416, "grad_norm": 0.00022003210324328393, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31824000.0, "reward": 1.4996765851974487, "reward_std": 0.0005541958380490541, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9993499517440796, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9862500429153442, "step": 2708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 319.3125, "completions/mean_terminated_length": 319.3125, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.5418, "grad_norm": 0.19467173516750336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31832997.0, "reward": 1.3217072486877441, "reward_std": 0.3808844983577728, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8309124708175659, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 388.6875, "completions/mean_terminated_length": 388.6875, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.542, "grad_norm": 0.14945538341999054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31842992.0, "reward": 0.7268759608268738, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26624998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 386.25, "completions/mean_terminated_length": 386.25, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.5422, "grad_norm": 0.11118496209383011, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31853076.0, "reward": 1.3457040786743164, "reward_std": 0.4300370216369629, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.87890625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 438.125, "completions/mean_terminated_length": 438.125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.5424, "grad_norm": 0.002845878480002284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31865302.0, "reward": 0.5469415187835693, "reward_std": 0.007237497717142105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09388124942779541, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 526.9375, "completions/mean_terminated_length": 526.9375, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.5426, "grad_norm": 0.03968673571944237, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31877573.0, "reward": 0.8391410708427429, "reward_std": 0.16058053076267242, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.678281307220459, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.49687501788139343, "step": 2713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 491.375, "completions/mean_terminated_length": 455.86669921875, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.5428, "grad_norm": 0.2017459273338318, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31890027.0, "reward": 1.283204197883606, "reward_std": 0.47629857063293457, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.81640625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 2714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 573.9375, "completions/mean_terminated_length": 573.9375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.543, "grad_norm": 0.0337560661137104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31904266.0, "reward": 0.6006947755813599, "reward_std": 0.2406049370765686, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13888749480247498, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 604.0, "completions/mean_length": 474.375, "completions/mean_terminated_length": 437.7333679199219, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.5432, "grad_norm": 0.14167173206806183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31915952.0, "reward": 0.9515600204467773, "reward_std": 0.4575329124927521, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6531187295913696, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6431249976158142, "step": 2716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 388.1875, "completions/mean_terminated_length": 388.1875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.5434, "grad_norm": 7.649084000149742e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31926195.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 340.875, "completions/mean_terminated_length": 340.875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.5436, "grad_norm": 0.1595936119556427, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31935553.0, "reward": 1.1997321844100952, "reward_std": 0.4587762951850891, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7119625210762024, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 642.5625, "completions/mean_terminated_length": 642.5625, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.5438, "grad_norm": 0.09404970705509186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31951050.0, "reward": 0.6955351829528809, "reward_std": 0.3273342549800873, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3910687267780304, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6768749952316284, "step": 2719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 456.75, "completions/mean_terminated_length": 418.933349609375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.544, "grad_norm": 0.11072877049446106, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31962086.0, "reward": 1.2866008281707764, "reward_std": 0.47124919295310974, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8232000470161438, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 2720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 440.1875, "completions/mean_terminated_length": 440.1875, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.5442, "grad_norm": 0.16234713792800903, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31973913.0, "reward": 0.8524229526519775, "reward_std": 0.45345523953437805, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39234375953674316, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 417.125, "completions/mean_terminated_length": 417.125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.5444, "grad_norm": 0.09668584913015366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31984731.0, "reward": 1.3800666332244873, "reward_std": 0.3245124816894531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 494.25, "completions/mean_terminated_length": 458.933349609375, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.5446, "grad_norm": 0.1046740710735321, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31996895.0, "reward": 0.6017196178436279, "reward_std": 0.3008737564086914, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20343749225139618, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 2723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 515.1875, "completions/mean_terminated_length": 515.1875, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.5448, "grad_norm": 0.09242472797632217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32009106.0, "reward": 0.6673448085784912, "reward_std": 0.32501113414764404, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2096875011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 293.6875, "completions/mean_terminated_length": 293.6875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.545, "grad_norm": 0.17184379696846008, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32017469.0, "reward": 0.49968576431274414, "reward_std": 0.13616997003555298, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06186874955892563, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 2725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 675.125, "completions/mean_terminated_length": 675.125, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.5452, "grad_norm": 0.05984947457909584, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32032015.0, "reward": 0.6311726570129395, "reward_std": 0.15893332660198212, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26234376430511475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.815625011920929, "step": 2726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 289.8125, "completions/mean_terminated_length": 289.8125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.5454, "grad_norm": 0.1312866359949112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32040924.0, "reward": 1.3198320865631104, "reward_std": 0.38257238268852234, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.827162504196167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 2727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 387.0625, "completions/mean_terminated_length": 387.0625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5456, "grad_norm": 0.1912512481212616, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32050445.0, "reward": 1.0206259489059448, "reward_std": 0.4880939722061157, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5412499904632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 2728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 997.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 796.375, "completions/mean_terminated_length": 796.375, "completions/min_length": 616.0, "completions/min_terminated_length": 616.0, "epoch": 0.5458, "grad_norm": 0.06591916084289551, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32067123.0, "reward": 0.7516412138938904, "reward_std": 0.20582200586795807, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5032812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5968749523162842, "step": 2729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 437.5, "completions/mean_terminated_length": 437.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.546, "grad_norm": 0.3183616101741791, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32079003.0, "reward": 0.8556264638900757, "reward_std": 0.5307857394218445, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3987500071525574, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 457.5, "completions/mean_terminated_length": 419.7333679199219, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.5462, "grad_norm": 0.14651863276958466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32090531.0, "reward": 1.1659009456634521, "reward_std": 0.5253235697746277, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7067999839782715, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 2731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 404.1875, "completions/mean_terminated_length": 404.1875, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.5464, "grad_norm": 0.18315927684307098, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32101238.0, "reward": 0.6656915545463562, "reward_std": 0.3252728283405304, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20638124644756317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 338.625, "completions/mean_terminated_length": 338.625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5466, "grad_norm": 0.24303098022937775, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32110848.0, "reward": 1.1427197456359863, "reward_std": 0.47517523169517517, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6604375243186951, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 347.8125, "completions/mean_terminated_length": 347.8125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.5468, "grad_norm": 0.14332707226276398, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32120141.0, "reward": 0.7148450613021851, "reward_std": 0.38990309834480286, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2421874850988388, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 2734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 395.125, "completions/mean_terminated_length": 395.125, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.547, "grad_norm": 0.11923924833536148, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32130655.0, "reward": 0.6137509346008301, "reward_std": 0.23810867965221405, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.28999999165534973, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "step": 2735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 692.6875, "completions/mean_terminated_length": 616.2307739257812, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.5472, "grad_norm": 0.14595189690589905, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32145418.0, "reward": 0.8284133076667786, "reward_std": 0.5929437875747681, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4693250060081482, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7287499904632568, "step": 2736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 696.1875, "completions/mean_terminated_length": 696.1875, "completions/min_length": 505.0, "completions/min_terminated_length": 505.0, "epoch": 0.5474, "grad_norm": 0.06954935193061829, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32163229.0, "reward": 0.8089818358421326, "reward_std": 0.21850159764289856, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6179624795913696, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4587499797344208, "step": 2737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 544.875, "completions/mean_terminated_length": 544.875, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.5476, "grad_norm": 0.04405933618545532, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32176667.0, "reward": 0.6101542711257935, "reward_std": 0.24279291927814484, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15780624747276306, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 2738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 537.0, "completions/max_terminated_length": 537.0, "completions/mean_length": 361.4375, "completions/mean_terminated_length": 361.4375, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.5478, "grad_norm": 0.17535801231861115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32186418.0, "reward": 1.0885134935379028, "reward_std": 0.4798153042793274, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6145249605178833, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999380111694, "step": 2739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 578.75, "completions/mean_terminated_length": 549.0667114257812, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.548, "grad_norm": 0.11453330516815186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32199822.0, "reward": 0.9865635633468628, "reward_std": 0.5453673005104065, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5356249809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 2740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 254.0, "completions/max_terminated_length": 254.0, "completions/mean_length": 182.875, "completions/mean_terminated_length": 182.875, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.5482, "grad_norm": 0.4447622299194336, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32206476.0, "reward": 1.1372541189193726, "reward_std": 0.4826558828353882, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6495062112808228, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631250500679016, "step": 2741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1017.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 705.875, "completions/mean_terminated_length": 705.875, "completions/min_length": 457.0, "completions/min_terminated_length": 457.0, "epoch": 0.5484, "grad_norm": 0.019091736525297165, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32221402.0, "reward": 0.9739846587181091, "reward_std": 0.10950563102960587, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.8854687213897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.22187501192092896, "step": 2742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 324.5625, "completions/mean_terminated_length": 324.5625, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.5486, "grad_norm": 0.0002097723918268457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32230643.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 461.0625, "completions/mean_terminated_length": 461.0625, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.5488, "grad_norm": 0.1138584166765213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32241908.0, "reward": 1.2623322010040283, "reward_std": 0.42314741015434265, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7746624946594238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 470.875, "completions/mean_terminated_length": 434.0000305175781, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.549, "grad_norm": 0.0505654402077198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32253602.0, "reward": 0.5735135078430176, "reward_std": 0.28487440943717957, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.147025004029274, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 2745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 482.75, "completions/mean_terminated_length": 482.75, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.5492, "grad_norm": 0.024458352476358414, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32265246.0, "reward": 0.5775009393692017, "reward_std": 0.08484035730361938, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1550000011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 2746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 377.5, "completions/mean_terminated_length": 377.5, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.5494, "grad_norm": 0.12022876739501953, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32275398.0, "reward": 0.8456915616989136, "reward_std": 0.45487844944000244, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3788812458515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 247.375, "completions/mean_terminated_length": 247.375, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.5496, "grad_norm": 0.007848801091313362, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32283260.0, "reward": 0.5085828900337219, "reward_std": 0.0162701103836298, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.017162499949336052, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9912500381469727, "step": 2748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 616.4375, "completions/mean_terminated_length": 558.2142944335938, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.5498, "grad_norm": 0.0628945603966713, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32296691.0, "reward": 0.53031325340271, "reward_std": 0.22817663848400116, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18562500178813934, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 2749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 600.0, "completions/mean_terminated_length": 539.4285888671875, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.55, "grad_norm": 0.11207077652215958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32309779.0, "reward": 0.7287504076957703, "reward_std": 0.3365492522716522, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5824999809265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.40625, "step": 2750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 542.375, "completions/mean_terminated_length": 542.375, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.5502, "grad_norm": 0.09687090665102005, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32322185.0, "reward": 0.7802821397781372, "reward_std": 0.4287777245044708, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31056249141693115, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 534.9375, "completions/mean_terminated_length": 534.9375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.5504, "grad_norm": 0.12673965096473694, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32334280.0, "reward": 0.7729573845863342, "reward_std": 0.38518771529197693, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.358412504196167, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8737499713897705, "step": 2752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.5506, "grad_norm": 0.1733601689338684, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32344029.0, "reward": 1.2836732864379883, "reward_std": 0.4765237867832184, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8173437118530273, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.909375011920929, "step": 2753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 324.0, "completions/mean_terminated_length": 324.0, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.5508, "grad_norm": 0.16782991588115692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32353469.0, "reward": 1.3190885782241821, "reward_std": 0.3870306611061096, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8256750106811523, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 2754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 740.6875, "completions/mean_terminated_length": 646.25, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.551, "grad_norm": 0.06672186404466629, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32370680.0, "reward": 0.6777348518371582, "reward_std": 0.4301489591598511, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.60546875, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.20937499403953552, "step": 2755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 786.25, "completions/mean_terminated_length": 770.4000244140625, "completions/min_length": 634.0, "completions/min_terminated_length": 634.0, "epoch": 0.5512, "grad_norm": 0.10334398597478867, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32387292.0, "reward": 0.7688947916030884, "reward_std": 0.47207486629486084, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3502874970436096, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8487499952316284, "step": 2756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 404.5625, "completions/mean_terminated_length": 404.5625, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.5514, "grad_norm": 0.03883841261267662, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32397749.0, "reward": 0.8698441982269287, "reward_std": 0.12433204799890518, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7396875023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 2757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 899.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 447.4375, "completions/mean_terminated_length": 447.4375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.5516, "grad_norm": 0.10622279345989227, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32409052.0, "reward": 0.8769289255142212, "reward_std": 0.366337388753891, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5663562417030334, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6818749904632568, "step": 2758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 375.4375, "completions/mean_terminated_length": 375.4375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5518, "grad_norm": 0.011003091931343079, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32418883.0, "reward": 0.543029248714447, "reward_std": 0.028627095744013786, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.086056187748909, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9555624723434448, "step": 2759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 1024.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 852.25, "completions/mean_terminated_length": 566.0, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.552, "grad_norm": 0.11204273253679276, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32437303.0, "reward": 0.33484405279159546, "reward_std": 0.4936872720718384, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23218749463558197, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.23125001788139343, "step": 2760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 451.5625, "completions/mean_terminated_length": 451.5625, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.5522, "grad_norm": 0.06792108714580536, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32449392.0, "reward": 1.4382292032241821, "reward_std": 0.24491487443447113, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9389562606811523, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 2761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 500.6875, "completions/mean_terminated_length": 500.6875, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.5524, "grad_norm": 0.11626700311899185, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32460891.0, "reward": 1.2542541027069092, "reward_std": 0.43814870715141296, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7585062980651855, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 565.4375, "completions/mean_terminated_length": 499.9285888671875, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.5526, "grad_norm": 0.06977032124996185, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32473954.0, "reward": 0.6035945415496826, "reward_std": 0.2970019280910492, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3321874737739563, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.625, "step": 2763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 347.9375, "completions/mean_terminated_length": 347.9375, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.5528, "grad_norm": 0.1474139243364334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32483249.0, "reward": 1.383044719696045, "reward_std": 0.31676071882247925, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.891087532043457, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 361.0, "completions/mean_terminated_length": 361.0, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.553, "grad_norm": 0.19527225196361542, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32493201.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 505.125, "completions/mean_terminated_length": 505.125, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.5532, "grad_norm": 0.0696980282664299, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32505059.0, "reward": 0.7517974972724915, "reward_std": 0.18137063086032867, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5035937428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.640625, "step": 2766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 837.0, "completions/mean_length": 730.125, "completions/mean_terminated_length": 688.1428833007812, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.5534, "grad_norm": 0.1398715078830719, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32520645.0, "reward": 0.7163883447647095, "reward_std": 0.501290500164032, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30777502059936523, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8349999785423279, "step": 2767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 842.6875, "completions/mean_terminated_length": 661.375, "completions/min_length": 485.0, "completions/min_terminated_length": 485.0, "epoch": 0.5536, "grad_norm": 0.05903509259223938, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32538096.0, "reward": 0.32951611280441284, "reward_std": 0.4118933081626892, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09653125703334808, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4793750047683716, "step": 2768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 623.5625, "completions/mean_terminated_length": 623.5625, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.5538, "grad_norm": 0.12322778254747391, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32553193.0, "reward": 0.6884102821350098, "reward_std": 0.42433398962020874, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25181877613067627, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.956250011920929, "step": 2769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 417.25, "completions/mean_terminated_length": 417.25, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.554, "grad_norm": 0.15445713698863983, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32564013.0, "reward": 1.2821884155273438, "reward_std": 0.38827547430992126, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8143750429153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 2770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 375.5625, "completions/mean_terminated_length": 375.5625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.5542, "grad_norm": 0.1807742863893509, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32573574.0, "reward": 1.259507179260254, "reward_std": 0.4281778037548065, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7690125107765198, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 617.0625, "completions/mean_terminated_length": 558.9285888671875, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.5544, "grad_norm": 0.15653492510318756, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32587639.0, "reward": 0.7711724042892456, "reward_std": 0.38552820682525635, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6048437356948853, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.37187498807907104, "step": 2772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 516.4375, "completions/mean_terminated_length": 516.4375, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.5546, "grad_norm": 0.1708756983280182, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32599950.0, "reward": 0.8739696741104126, "reward_std": 0.4493052661418915, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.43543750047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8949999809265137, "step": 2773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 415.875, "completions/mean_terminated_length": 415.875, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.5548, "grad_norm": 0.14848096668720245, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32610332.0, "reward": 0.9659383296966553, "reward_std": 0.40445828437805176, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6193749904632568, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 2774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 328.375, "completions/mean_terminated_length": 328.375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.555, "grad_norm": 0.17350740730762482, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32619714.0, "reward": 0.7846100330352783, "reward_std": 0.27620840072631836, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5067187547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6468749642372131, "step": 2775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 752.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 607.25, "completions/mean_terminated_length": 607.25, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.5552, "grad_norm": 0.040560729801654816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32632886.0, "reward": 0.5677324533462524, "reward_std": 0.11180683225393295, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13546249270439148, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9149999618530273, "step": 2776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 419.25, "completions/mean_terminated_length": 419.25, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.5554, "grad_norm": 0.007554436102509499, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32643786.0, "reward": 0.5317636132240295, "reward_std": 0.02053987607359886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0635249987244606, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 2777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 455.25, "completions/mean_terminated_length": 455.25, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.5556, "grad_norm": 0.12667760252952576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32654782.0, "reward": 1.3487385511398315, "reward_std": 0.3365223705768585, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.884975016117096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 2778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 360.0, "completions/max_terminated_length": 360.0, "completions/mean_length": 291.625, "completions/mean_terminated_length": 291.625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.5558, "grad_norm": 0.06490800529718399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32664024.0, "reward": 1.4391415119171143, "reward_std": 0.2374393194913864, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9407812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 2779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 455.75, "completions/mean_terminated_length": 455.75, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.556, "grad_norm": 0.11256149411201477, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32675268.0, "reward": 1.2542072534561157, "reward_std": 0.4380909502506256, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7584124803543091, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 2780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 261.625, "completions/mean_terminated_length": 261.625, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.5562, "grad_norm": 0.13543492555618286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32683342.0, "reward": 1.2750539779663086, "reward_std": 0.33863139152526855, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8626062870025635, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7881250381469727, "step": 2781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 357.1875, "completions/mean_terminated_length": 357.1875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.5564, "grad_norm": 0.03871865198016167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32693169.0, "reward": 1.4659727811813354, "reward_std": 0.13259373605251312, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9944437742233276, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9168750047683716, "step": 2782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 353.9375, "completions/mean_terminated_length": 309.2666931152344, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.5566, "grad_norm": 0.21987611055374146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32702608.0, "reward": 1.013985276222229, "reward_std": 0.49161553382873535, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6529687643051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 2783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 400.75, "completions/mean_terminated_length": 400.75, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.5568, "grad_norm": 0.08660652488470078, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32713308.0, "reward": 1.439516544342041, "reward_std": 0.23753780126571655, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.941531240940094, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 638.875, "completions/mean_terminated_length": 613.2000122070312, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.557, "grad_norm": 0.07207953929901123, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32727146.0, "reward": 0.682266354560852, "reward_std": 0.2898711562156677, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3645312488079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7468749284744263, "step": 2785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 295.4375, "completions/mean_terminated_length": 295.4375, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.5572, "grad_norm": 0.16310876607894897, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32735681.0, "reward": 0.715929388999939, "reward_std": 0.3896375596523285, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24435624480247498, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 2786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 591.875, "completions/mean_terminated_length": 591.875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.5574, "grad_norm": 0.13416554033756256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32748991.0, "reward": 1.072188377380371, "reward_std": 0.40003639459609985, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7068750262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 2787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 394.25, "completions/mean_terminated_length": 394.25, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.5576, "grad_norm": 0.15954077243804932, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32758723.0, "reward": 0.7860942482948303, "reward_std": 0.24755236506462097, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6346875429153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.53125, "step": 2788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 543.0, "completions/mean_terminated_length": 543.0, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.5578, "grad_norm": 0.09168489277362823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32772403.0, "reward": 0.7772541046142578, "reward_std": 0.32090187072753906, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.42950624227523804, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7706249952316284, "step": 2789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 873.5, "completions/mean_terminated_length": 805.0909423828125, "completions/min_length": 606.0, "completions/min_terminated_length": 606.0, "epoch": 0.558, "grad_norm": 0.1645219922065735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32790555.0, "reward": 0.8496097326278687, "reward_std": 0.6336628794670105, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.38671875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.34062498807907104, "step": 2790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 464.5625, "completions/mean_terminated_length": 464.5625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.5582, "grad_norm": 0.19055791199207306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32801764.0, "reward": 1.1733227968215942, "reward_std": 0.4458020031452179, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7216437458992004, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8868750333786011, "step": 2791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 456.0, "completions/mean_terminated_length": 456.0, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.5584, "grad_norm": 0.0612141452729702, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32812724.0, "reward": 0.71976637840271, "reward_std": 0.26993393898010254, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.37703126668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.778124988079071, "step": 2792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 795.75, "completions/mean_terminated_length": 692.0, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.5586, "grad_norm": 0.12841111421585083, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32829520.0, "reward": 0.5169411897659302, "reward_std": 0.4448494017124176, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2838812470436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5018750429153442, "step": 2793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 688.8125, "completions/mean_terminated_length": 688.8125, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.5588, "grad_norm": 0.14980781078338623, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32845373.0, "reward": 1.0652353763580322, "reward_std": 0.5255444645881653, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6304687261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 2794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 740.1875, "completions/mean_terminated_length": 699.6428833007812, "completions/min_length": 606.0, "completions/min_terminated_length": 606.0, "epoch": 0.559, "grad_norm": 0.11000797152519226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32861392.0, "reward": 0.9175259470939636, "reward_std": 0.5749080181121826, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5225499868392944, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7850000262260437, "step": 2795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 299.6875, "completions/mean_terminated_length": 299.6875, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.5592, "grad_norm": 0.1338980346918106, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32870491.0, "reward": 0.7190573215484619, "reward_std": 0.38740137219429016, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2506124973297119, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 376.8125, "completions/mean_terminated_length": 376.8125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.5594, "grad_norm": 0.1301746368408203, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32880440.0, "reward": 1.067251443862915, "reward_std": 0.5065646171569824, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5720000267028809, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.987500011920929, "step": 2797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 382.625, "completions/mean_terminated_length": 339.86669921875, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5596, "grad_norm": 0.0833829939365387, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32890386.0, "reward": 1.3460415601730347, "reward_std": 0.4301656484603882, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8795812726020813, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9006249904632568, "step": 2798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 554.9375, "completions/mean_terminated_length": 554.9375, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.5598, "grad_norm": 0.07970020174980164, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32903425.0, "reward": 0.6049230098724365, "reward_std": 0.23929162323474884, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14734375476837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 610.8125, "completions/mean_terminated_length": 583.2667236328125, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.56, "grad_norm": 0.11487747728824615, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32917758.0, "reward": 0.786719560623169, "reward_std": 0.4540484845638275, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38593751192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 2800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 265.5, "completions/mean_terminated_length": 265.5, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.5602, "grad_norm": 0.0003186744579579681, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32926230.0, "reward": 1.499297857284546, "reward_std": 0.0005676754517480731, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985937476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 538.6875, "completions/mean_terminated_length": 538.6875, "completions/min_length": 478.0, "completions/min_terminated_length": 478.0, "epoch": 0.5604, "grad_norm": 0.09555109590291977, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32939249.0, "reward": 1.1562011241912842, "reward_std": 0.46548885107040405, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6873999834060669, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9500000476837158, "step": 2802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 306.5, "completions/mean_terminated_length": 306.5, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.5606, "grad_norm": 0.18756234645843506, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32947609.0, "reward": 0.8984760046005249, "reward_std": 0.48072224855422974, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42195001244544983, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 2803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 441.625, "completions/mean_terminated_length": 402.8000183105469, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5608, "grad_norm": 0.16322040557861328, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32958499.0, "reward": 0.7496072053909302, "reward_std": 0.46687039732933044, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31171250343322754, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 2804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 362.1875, "completions/mean_terminated_length": 362.1875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.561, "grad_norm": 0.17999328672885895, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32967910.0, "reward": 1.2620198726654053, "reward_std": 0.42448627948760986, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7740374803543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 2805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 964.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 644.3125, "completions/mean_terminated_length": 644.3125, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.5612, "grad_norm": 0.04155559092760086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32983131.0, "reward": 0.9496097564697266, "reward_std": 0.10740066319704056, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8992187976837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.20937500894069672, "step": 2806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 513.8125, "completions/mean_terminated_length": 513.8125, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.5614, "grad_norm": 0.16832545399665833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32995784.0, "reward": 1.0353920459747314, "reward_std": 0.48691868782043457, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5707812309265137, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468750357627869, "step": 2807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 379.0, "completions/mean_terminated_length": 379.0, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.5616, "grad_norm": 0.003726457478478551, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33006024.0, "reward": 0.5433228015899658, "reward_std": 0.011670105159282684, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08664374053478241, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 804.375, "completions/mean_terminated_length": 731.1666870117188, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.5618, "grad_norm": 0.13074038922786713, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33023118.0, "reward": 0.740769624710083, "reward_std": 0.5953643321990967, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41903749108314514, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6737499833106995, "step": 2809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 520.4375, "completions/mean_terminated_length": 486.86669921875, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.562, "grad_norm": 0.12225009500980377, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33035221.0, "reward": 1.2269539833068848, "reward_std": 0.501449704170227, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7664062976837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 507.875, "completions/mean_terminated_length": 507.875, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5622, "grad_norm": 0.14826519787311554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33048227.0, "reward": 0.7773852348327637, "reward_std": 0.3858186900615692, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.36726874113082886, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 2811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 378.75, "completions/mean_terminated_length": 378.75, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5624, "grad_norm": 0.11212874203920364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33057951.0, "reward": 1.3784854412078857, "reward_std": 0.329588383436203, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8819687366485596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 315.0, "completions/max_terminated_length": 315.0, "completions/mean_length": 245.6875, "completions/mean_terminated_length": 245.6875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.5626, "grad_norm": 0.21359768509864807, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33065754.0, "reward": 0.8827792406082153, "reward_std": 0.49336040019989014, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3905562460422516, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 2813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 442.0, "completions/mean_terminated_length": 442.0, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.5628, "grad_norm": 0.07115507125854492, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33076106.0, "reward": 0.5991165637969971, "reward_std": 0.24076944589614868, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13573125004768372, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 2814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 394.125, "completions/mean_terminated_length": 394.125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.563, "grad_norm": 0.20255129039287567, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33086252.0, "reward": 1.2863945960998535, "reward_std": 0.47114965319633484, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8227875232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 2815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 489.4375, "completions/mean_terminated_length": 489.4375, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.5632, "grad_norm": 0.17975987493991852, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33098499.0, "reward": 0.9188290238380432, "reward_std": 0.46540313959121704, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.46265625953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 2816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 756.0, "completions/mean_terminated_length": 756.0, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.5634, "grad_norm": 0.01295996829867363, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33114451.0, "reward": 0.5643759965896606, "reward_std": 0.054946813732385635, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1287499964237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 2817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 528.0, "completions/mean_terminated_length": 494.933349609375, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.5636, "grad_norm": 0.07151518017053604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33127475.0, "reward": 0.9380471706390381, "reward_std": 0.28705304861068726, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.8760937452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.24687498807907104, "step": 2818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 826.0, "completions/mean_length": 674.3125, "completions/mean_terminated_length": 624.357177734375, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.5638, "grad_norm": 0.043698500841856, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33142760.0, "reward": 0.5154664516448975, "reward_std": 0.23524294793605804, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1559312492609024, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.768125057220459, "step": 2819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 361.1875, "completions/mean_terminated_length": 361.1875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.564, "grad_norm": 0.00014555465895682573, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33153163.0, "reward": 1.4989008903503418, "reward_std": 0.0004099227662663907, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977999925613403, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 347.25, "completions/mean_terminated_length": 347.25, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5642, "grad_norm": 0.11530745774507523, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33163519.0, "reward": 1.261394739151001, "reward_std": 0.424938827753067, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7727875113487244, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 2821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 390.125, "completions/mean_terminated_length": 390.125, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.5644, "grad_norm": 0.00849597156047821, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33173873.0, "reward": 0.5551446676254272, "reward_std": 0.03406907618045807, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11028750240802765, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 2822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 235.9375, "completions/mean_terminated_length": 235.9375, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.5646, "grad_norm": 0.19472919404506683, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33181632.0, "reward": 1.341076135635376, "reward_std": 0.33380600810050964, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8696500062942505, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8787500262260437, "step": 2823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 585.3125, "completions/mean_terminated_length": 585.3125, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.5648, "grad_norm": 0.13134536147117615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33194469.0, "reward": 0.8395822048187256, "reward_std": 0.45915549993515015, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36666250228881836, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 2824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 368.4375, "completions/mean_terminated_length": 368.4375, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.565, "grad_norm": 0.07329050451517105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33203916.0, "reward": 1.4366885423660278, "reward_std": 0.24978366494178772, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9358749389648438, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9649375081062317, "step": 2825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 296.25, "completions/mean_terminated_length": 296.25, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.5652, "grad_norm": 0.07840842753648758, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33213184.0, "reward": 1.4400792121887207, "reward_std": 0.23768816888332367, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9426562786102295, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 2826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 698.25, "completions/mean_terminated_length": 676.5333862304688, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.5654, "grad_norm": 0.07320600748062134, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33228228.0, "reward": 0.6766412258148193, "reward_std": 0.24341049790382385, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.41578125953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.659375011920929, "step": 2827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 455.8125, "completions/mean_terminated_length": 455.8125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.5656, "grad_norm": 0.13789932429790497, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33239297.0, "reward": 0.7144542932510376, "reward_std": 0.3890649080276489, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24140623211860657, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 2828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 418.5, "completions/mean_terminated_length": 418.5, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.5658, "grad_norm": 0.13823865354061127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33250041.0, "reward": 1.261394739151001, "reward_std": 0.424938827753067, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7727875113487244, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 311.0, "completions/max_terminated_length": 311.0, "completions/mean_length": 230.6875, "completions/mean_terminated_length": 230.6875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.566, "grad_norm": 0.00037485628854483366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33257876.0, "reward": 1.4993228912353516, "reward_std": 0.0005930570769123733, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986437559127808, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 2830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 342.3125, "completions/mean_terminated_length": 342.3125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.5662, "grad_norm": 0.18255585432052612, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33267529.0, "reward": 1.0238292217254639, "reward_std": 0.49128803610801697, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.547656238079071, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 2831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 268.0625, "completions/mean_terminated_length": 268.0625, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.5664, "grad_norm": 0.11067088693380356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33276778.0, "reward": 1.4485697746276855, "reward_std": 0.1852002888917923, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.959637463092804, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 2832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 576.625, "completions/mean_terminated_length": 576.625, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5666, "grad_norm": 0.10106754302978516, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33290132.0, "reward": 0.7979697585105896, "reward_std": 0.29677221179008484, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5334374904632568, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.606249988079071, "step": 2833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 229.0, "completions/max_terminated_length": 229.0, "completions/mean_length": 187.5625, "completions/mean_terminated_length": 187.5625, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.5668, "grad_norm": 0.4742925465106964, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33296877.0, "reward": 1.1975511312484741, "reward_std": 0.4627036154270172, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7075999975204468, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 2834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 326.1875, "completions/mean_terminated_length": 326.1875, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.567, "grad_norm": 0.1897377371788025, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33306032.0, "reward": 1.320469856262207, "reward_std": 0.38368529081344604, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8284374475479126, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 417.8125, "completions/mean_terminated_length": 417.8125, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.5672, "grad_norm": 0.1858450472354889, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33316381.0, "reward": 0.9333603382110596, "reward_std": 0.46316054463386536, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49171873927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 2836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 495.9375, "completions/mean_terminated_length": 495.9375, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.5674, "grad_norm": 0.12383254617452621, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33328572.0, "reward": 1.1528820991516113, "reward_std": 0.42218074202537537, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7432624697685242, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8574999570846558, "step": 2837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 373.5, "completions/mean_terminated_length": 330.13336181640625, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.5676, "grad_norm": 0.3151003122329712, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33338244.0, "reward": 0.7521100640296936, "reward_std": 0.3747681975364685, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4417187571525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.703125, "step": 2838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 443.6875, "completions/mean_terminated_length": 443.6875, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.5678, "grad_norm": 0.008615010417997837, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33349103.0, "reward": 0.5572665929794312, "reward_std": 0.02460188791155815, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1145312488079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 543.1875, "completions/mean_terminated_length": 543.1875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.568, "grad_norm": 0.1647130399942398, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33362034.0, "reward": 0.7349227070808411, "reward_std": 0.28179097175598145, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4073437452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.715624988079071, "step": 2840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 502.5, "completions/mean_terminated_length": 502.5, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.5682, "grad_norm": 0.12494013458490372, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33373818.0, "reward": 0.7220259308815002, "reward_std": 0.3855838179588318, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25654998421669006, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 444.1875, "completions/mean_terminated_length": 405.5333557128906, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.5684, "grad_norm": 0.08403240144252777, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33384589.0, "reward": 0.8266410231590271, "reward_std": 0.26614826917648315, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7157812118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.37812501192092896, "step": 2842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 518.9375, "completions/mean_terminated_length": 485.2666931152344, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.5686, "grad_norm": 0.11042055487632751, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33397484.0, "reward": 0.7346881031990051, "reward_std": 0.29463934898376465, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.46937501430511475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5187499523162842, "step": 2843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 509.1875, "completions/mean_terminated_length": 474.86669921875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.5688, "grad_norm": 0.024276643991470337, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33410415.0, "reward": 0.513829231262207, "reward_std": 0.1411815881729126, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09015624970197678, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 2844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 737.125, "completions/mean_terminated_length": 641.5, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.569, "grad_norm": 0.0549749992787838, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33426241.0, "reward": 0.38773512840270996, "reward_std": 0.2737681567668915, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08796874433755875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.699999988079071, "step": 2845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 722.5625, "completions/mean_terminated_length": 653.0, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.5692, "grad_norm": 0.11174202710390091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33441546.0, "reward": 0.6603912115097046, "reward_std": 0.45361819863319397, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.38328123092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 2846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 459.8125, "completions/mean_terminated_length": 459.8125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5694, "grad_norm": 0.12348446249961853, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33452823.0, "reward": 1.100441575050354, "reward_std": 0.47333401441574097, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6383812427520752, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 2847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 285.625, "completions/mean_terminated_length": 285.625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.5696, "grad_norm": 0.1958560049533844, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33460977.0, "reward": 1.0143448114395142, "reward_std": 0.5010999441146851, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5286874771118164, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 2848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 466.25, "completions/mean_terminated_length": 466.25, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.5698, "grad_norm": 0.11252176761627197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33472501.0, "reward": 1.2311571836471558, "reward_std": 0.3823109567165375, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8373124599456787, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7699999809265137, "step": 2849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 453.125, "completions/mean_terminated_length": 453.125, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.57, "grad_norm": 0.07714535295963287, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33483767.0, "reward": 1.3764071464538574, "reward_std": 0.2793494164943695, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9403125047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 2850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 534.25, "completions/mean_terminated_length": 501.60003662109375, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.5702, "grad_norm": 0.14382991194725037, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33496043.0, "reward": 1.1047290563583374, "reward_std": 0.5407643914222717, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.646956205368042, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 2851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 412.3125, "completions/mean_terminated_length": 412.3125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.5704, "grad_norm": 0.04855313524603844, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33506416.0, "reward": 1.422579050064087, "reward_std": 0.1832825392484665, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9701562523841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 2852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 321.5625, "completions/mean_terminated_length": 321.5625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.5706, "grad_norm": 0.1852777749300003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33515657.0, "reward": 1.0831259489059448, "reward_std": 0.48673015832901, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6037499904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 590.6875, "completions/mean_terminated_length": 590.6875, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.5708, "grad_norm": 0.11391149461269379, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33528932.0, "reward": 1.3211040496826172, "reward_std": 0.38319480419158936, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8297062516212463, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 390.0, "completions/mean_terminated_length": 390.0, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.571, "grad_norm": 0.09894435107707977, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33539236.0, "reward": 1.3768634796142578, "reward_std": 0.33289751410484314, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.878724992275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 771.1875, "completions/mean_terminated_length": 735.0714721679688, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.5712, "grad_norm": 0.07021844387054443, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33557943.0, "reward": 0.6139727234840393, "reward_std": 0.377998948097229, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22794374823570251, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7393749952316284, "step": 2856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 440.9375, "completions/mean_terminated_length": 440.9375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.5714, "grad_norm": 0.14103014767169952, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33569046.0, "reward": 0.9079697132110596, "reward_std": 0.4729275405406952, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44093751907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 245.8125, "completions/mean_terminated_length": 245.8125, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.5716, "grad_norm": 0.19369496405124664, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33577011.0, "reward": 1.317879319190979, "reward_std": 0.38979408144950867, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.823256254196167, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 2858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 703.5, "completions/mean_terminated_length": 657.7142944335938, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.5718, "grad_norm": 0.08094324171543121, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33592731.0, "reward": 0.7241411209106445, "reward_std": 0.31906959414482117, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5732812285423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4468750059604645, "step": 2859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 255.0, "completions/max_terminated_length": 255.0, "completions/mean_length": 205.0625, "completions/mean_terminated_length": 205.0625, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.572, "grad_norm": 0.3241475522518158, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33600444.0, "reward": 1.437572956085205, "reward_std": 0.24736645817756653, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9376437664031982, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9718750715255737, "step": 2860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 177.0, "completions/max_terminated_length": 177.0, "completions/mean_length": 145.5625, "completions/mean_terminated_length": 145.5625, "completions/min_length": 118.0, "completions/min_terminated_length": 118.0, "epoch": 0.5722, "grad_norm": 0.00034333058283664286, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33606533.0, "reward": 1.499863624572754, "reward_std": 0.00030200224136933684, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9997249841690063, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.987500011920929, "step": 2861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 311.0, "completions/max_terminated_length": 311.0, "completions/mean_length": 241.5, "completions/mean_terminated_length": 241.5, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.5724, "grad_norm": 0.0825193002820015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33614205.0, "reward": 1.439507246017456, "reward_std": 0.2375352531671524, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 389.8125, "completions/mean_terminated_length": 389.8125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.5726, "grad_norm": 0.18145941197872162, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33625130.0, "reward": 1.000941514968872, "reward_std": 0.4472712576389313, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5643812417984009, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8056187629699707, "step": 2863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 372.9375, "completions/mean_terminated_length": 372.9375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.5728, "grad_norm": 0.12398206442594528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33634649.0, "reward": 1.3182759284973145, "reward_std": 0.3882943093776703, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8240500092506409, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 2864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 371.8125, "completions/mean_terminated_length": 371.8125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.573, "grad_norm": 0.05800577253103256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33645334.0, "reward": 1.439507246017456, "reward_std": 0.2375352680683136, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 342.375, "completions/mean_terminated_length": 342.375, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.5732, "grad_norm": 0.00016191115719266236, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33655180.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 760.6875, "completions/mean_terminated_length": 743.1333618164062, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.5734, "grad_norm": 0.053454454988241196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33671703.0, "reward": 0.7307037711143494, "reward_std": 0.3187190890312195, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.46140623092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.640625, "step": 2867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 552.0625, "completions/mean_terminated_length": 552.0625, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.5736, "grad_norm": 0.03242727741599083, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33684232.0, "reward": 0.7910162210464478, "reward_std": 0.12881183624267578, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.58203125, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 2868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 508.0625, "completions/mean_terminated_length": 508.0625, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.5738, "grad_norm": 0.013159891590476036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33696297.0, "reward": 0.5727353096008301, "reward_std": 0.05927368998527527, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14546874165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 2869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 484.25, "completions/mean_terminated_length": 484.25, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.574, "grad_norm": 0.14000122249126434, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33707773.0, "reward": 1.1936134099960327, "reward_std": 0.46796709299087524, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6997250318527222, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 2870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 871.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 522.625, "completions/mean_terminated_length": 522.625, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.5742, "grad_norm": 0.10590758919715881, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33720007.0, "reward": 0.954997718334198, "reward_std": 0.44732820987701416, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5349937677383423, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8618749976158142, "step": 2871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 524.75, "completions/mean_terminated_length": 491.4667053222656, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.5744, "grad_norm": 0.10789426416158676, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33732275.0, "reward": 1.1609076261520386, "reward_std": 0.5326740145683289, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6968125104904175, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 2872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 828.375, "completions/mean_terminated_length": 783.2307739257812, "completions/min_length": 602.0, "completions/min_terminated_length": 602.0, "epoch": 0.5746, "grad_norm": 0.05785553902387619, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33749689.0, "reward": 0.6687505841255188, "reward_std": 0.3986212909221649, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4625000059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.53125, "step": 2873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 530.375, "completions/mean_terminated_length": 497.4667053222656, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.5748, "grad_norm": 0.18544486165046692, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33762287.0, "reward": 0.9128664135932922, "reward_std": 0.5070984363555908, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5132312774658203, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8256250023841858, "step": 2874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 338.0, "completions/max_terminated_length": 338.0, "completions/mean_length": 294.375, "completions/mean_terminated_length": 294.375, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.575, "grad_norm": 0.08059507608413696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33771253.0, "reward": 0.605847954750061, "reward_std": 0.24000413715839386, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14919374883174896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 316.125, "completions/mean_terminated_length": 316.125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.5752, "grad_norm": 0.16760984063148499, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33780455.0, "reward": 0.8330979943275452, "reward_std": 0.46434861421585083, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35369375348091125, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 2876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 896.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 523.5625, "completions/mean_terminated_length": 523.5625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5754, "grad_norm": 0.12994039058685303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33793664.0, "reward": 1.1624195575714111, "reward_std": 0.37812814116477966, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.8248374462127686, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7024999856948853, "step": 2877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 492.8125, "completions/mean_terminated_length": 492.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.5756, "grad_norm": 0.1368117332458496, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33805677.0, "reward": 1.1102101802825928, "reward_std": 0.4668106436729431, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6579187512397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 2878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 293.0, "completions/max_terminated_length": 293.0, "completions/mean_length": 195.5625, "completions/mean_terminated_length": 195.5625, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 0.5758, "grad_norm": 0.19574667513370514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33812326.0, "reward": 1.439666509628296, "reward_std": 0.23758070170879364, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9418312311172485, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 2879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 370.25, "completions/mean_terminated_length": 370.25, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.576, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33822170.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 391.8125, "completions/mean_terminated_length": 391.8125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.5762, "grad_norm": 0.17656774818897247, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33831975.0, "reward": 1.201951026916504, "reward_std": 0.45483022928237915, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7164000272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 877.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 593.75, "completions/mean_terminated_length": 593.75, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.5764, "grad_norm": 0.09354270249605179, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33846675.0, "reward": 0.7232571840286255, "reward_std": 0.38487550616264343, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25901249051094055, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 218.0, "completions/max_terminated_length": 218.0, "completions/mean_length": 176.3125, "completions/mean_terminated_length": 176.3125, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "epoch": 0.5766, "grad_norm": 0.1399964690208435, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33853288.0, "reward": 1.3782541751861572, "reward_std": 0.3314540684223175, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8815062046051025, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 2883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 423.5625, "completions/mean_terminated_length": 423.5625, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.5768, "grad_norm": 0.15865936875343323, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33864257.0, "reward": 1.0847634077072144, "reward_std": 0.44326159358024597, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6695250272750854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8525000214576721, "step": 2884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 569.625, "completions/mean_terminated_length": 539.3333740234375, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.577, "grad_norm": 0.21005405485630035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33877291.0, "reward": 0.9925540089607239, "reward_std": 0.540262758731842, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5476062297821045, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 2885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 463.25, "completions/mean_terminated_length": 463.25, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.5772, "grad_norm": 0.07559019327163696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33890287.0, "reward": 0.6560947299003601, "reward_std": 0.3294772803783417, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18718749284744263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 2886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 600.25, "completions/mean_terminated_length": 572.0, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.5774, "grad_norm": 0.11660049110651016, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33904083.0, "reward": 0.7153913378715515, "reward_std": 0.28124961256980896, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4932812452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.534375011920929, "step": 2887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 222.0, "completions/max_terminated_length": 222.0, "completions/mean_length": 181.6875, "completions/mean_terminated_length": 181.6875, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.5776, "grad_norm": 0.1977681964635849, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33910302.0, "reward": 1.3760573863983154, "reward_std": 0.3380935788154602, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8771125078201294, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9862500429153442, "step": 2888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 671.3125, "completions/mean_terminated_length": 620.9285888671875, "completions/min_length": 450.0, "completions/min_terminated_length": 450.0, "epoch": 0.5778, "grad_norm": 0.0547754131257534, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33925491.0, "reward": 0.7752348184585571, "reward_std": 0.335175484418869, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6754687428474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3031249940395355, "step": 2889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 440.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 372.4375, "completions/mean_terminated_length": 372.4375, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.578, "grad_norm": 0.14173199236392975, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33935194.0, "reward": 1.200219750404358, "reward_std": 0.4577052593231201, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7129374742507935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 2890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 303.0, "completions/max_terminated_length": 303.0, "completions/mean_length": 227.4375, "completions/mean_terminated_length": 227.4375, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.5782, "grad_norm": 0.2334391474723816, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33942529.0, "reward": 1.2554854154586792, "reward_std": 0.43737557530403137, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7609687447547913, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.986875057220459, "step": 2891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 800.5, "completions/mean_terminated_length": 726.0, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.5784, "grad_norm": 0.1056181937456131, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33960009.0, "reward": 0.47937560081481934, "reward_std": 0.3806252181529999, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20874999463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5812499523162842, "step": 2892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 426.125, "completions/mean_terminated_length": 426.125, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.5786, "grad_norm": 0.005905103404074907, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33971547.0, "reward": 0.5485947728157043, "reward_std": 0.01734958030283451, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 534.8125, "completions/mean_terminated_length": 534.8125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.5788, "grad_norm": 0.07120224088430405, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33984120.0, "reward": 0.8648443222045898, "reward_std": 0.16133847832679749, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.729687511920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3999999761581421, "step": 2894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 431.0625, "completions/mean_terminated_length": 431.0625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.579, "grad_norm": 0.035383857786655426, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33995753.0, "reward": 0.969687819480896, "reward_std": 0.10000763833522797, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9393750429153442, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.14374999701976776, "step": 2895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 457.75, "completions/mean_terminated_length": 457.75, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.5792, "grad_norm": 0.11667689681053162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34006997.0, "reward": 1.2217977046966553, "reward_std": 0.4146459102630615, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.6935937404632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.840624988079071, "step": 2896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 429.4375, "completions/mean_terminated_length": 429.4375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.5794, "grad_norm": 0.04435121268033981, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34017436.0, "reward": 0.6567195653915405, "reward_std": 0.1444108784198761, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3134375214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 2897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 901.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 623.25, "completions/mean_terminated_length": 623.25, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.5796, "grad_norm": 0.08122339099645615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34031664.0, "reward": 0.6873446702957153, "reward_std": 0.26668038964271545, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3121874928474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 2898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 699.8125, "completions/mean_terminated_length": 699.8125, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.5798, "grad_norm": 0.08152266591787338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34046397.0, "reward": 0.8085167407989502, "reward_std": 0.424629807472229, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3670312464237213, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 2899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 407.0, "completions/max_terminated_length": 407.0, "completions/mean_length": 295.125, "completions/mean_terminated_length": 295.125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.58, "grad_norm": 0.19548411667346954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34055423.0, "reward": 1.0219415426254272, "reward_std": 0.4924962520599365, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5438812971115112, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 2900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 330.5625, "completions/mean_terminated_length": 330.5625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.5802, "grad_norm": 0.16410735249519348, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34064808.0, "reward": 0.9573854207992554, "reward_std": 0.4939194917678833, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4772687554359436, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 240.4375, "completions/mean_terminated_length": 240.4375, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.5804, "grad_norm": 0.21820227801799774, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34072415.0, "reward": 1.207650899887085, "reward_std": 0.44674941897392273, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7278000116348267, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 422.5, "completions/mean_terminated_length": 422.5, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.5806, "grad_norm": 0.1316271871328354, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34083191.0, "reward": 0.9281134009361267, "reward_std": 0.4630196690559387, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.48122501373291016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 2903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 335.1875, "completions/mean_terminated_length": 335.1875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.5808, "grad_norm": 0.0017489552265033126, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34092186.0, "reward": 1.497175931930542, "reward_std": 0.007598598022013903, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.994350016117096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 2904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 408.6875, "completions/mean_terminated_length": 408.6875, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.581, "grad_norm": 0.17044030129909515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34102709.0, "reward": 1.081382155418396, "reward_std": 0.48895788192749023, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6002625226974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 261.4375, "completions/mean_terminated_length": 261.4375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.5812, "grad_norm": 0.15859930217266083, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34110636.0, "reward": 1.3461666107177734, "reward_std": 0.33942654728889465, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8798313140869141, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143750667572021, "step": 2906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 274.625, "completions/mean_terminated_length": 274.625, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.5814, "grad_norm": 0.11026456952095032, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34118950.0, "reward": 0.588204026222229, "reward_std": 0.15781231224536896, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11390624940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 2907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 445.3125, "completions/mean_terminated_length": 445.3125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.5816, "grad_norm": 0.13478268682956696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34130203.0, "reward": 0.9670320749282837, "reward_std": 0.4839545488357544, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49656248092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 315.0, "completions/max_terminated_length": 315.0, "completions/mean_length": 258.375, "completions/mean_terminated_length": 258.375, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.5818, "grad_norm": 0.0002632733667269349, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34138689.0, "reward": 1.4994856119155884, "reward_std": 0.0005917706876061857, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989687204360962, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 2909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 406.4375, "completions/mean_terminated_length": 406.4375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.582, "grad_norm": 0.17914517223834991, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34148712.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 772.3125, "completions/mean_terminated_length": 714.2307739257812, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.5822, "grad_norm": 0.08238938450813293, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34165069.0, "reward": 0.6212507486343384, "reward_std": 0.44392871856689453, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3050000071525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.675000011920929, "step": 2911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 348.625, "completions/mean_terminated_length": 348.625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.5824, "grad_norm": 0.11150472611188889, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34174567.0, "reward": 1.3801541328430176, "reward_std": 0.32454684376716614, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8853062391281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 568.5, "completions/mean_terminated_length": 538.1333618164062, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.5826, "grad_norm": 0.11415812373161316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34188415.0, "reward": 0.7270322442054749, "reward_std": 0.34657299518585205, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.45406249165534973, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5874999761581421, "step": 2913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 538.6875, "completions/mean_terminated_length": 538.6875, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.5828, "grad_norm": 0.08831702172756195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34201834.0, "reward": 0.8002350330352783, "reward_std": 0.23004767298698425, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6004687547683716, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4593750238418579, "step": 2914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 934.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 693.875, "completions/mean_terminated_length": 693.875, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.583, "grad_norm": 0.18816211819648743, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34216552.0, "reward": 1.4004571437835693, "reward_std": 0.27531546354293823, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9259124994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 2915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 483.0, "completions/mean_terminated_length": 483.0, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.5832, "grad_norm": 0.11671556532382965, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34228952.0, "reward": 0.856329083442688, "reward_std": 0.44945216178894043, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.40015625953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 2916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 327.375, "completions/mean_terminated_length": 327.375, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.5834, "grad_norm": 0.10271281749010086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34237870.0, "reward": 1.2610821723937988, "reward_std": 0.42475342750549316, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7721624970436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 531.75, "completions/mean_terminated_length": 498.933349609375, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.5836, "grad_norm": 0.14984844624996185, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34250234.0, "reward": 0.9839725494384766, "reward_std": 0.43261611461639404, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7179437875747681, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6206250190734863, "step": 2918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 471.875, "completions/mean_terminated_length": 393.0000305175781, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.5838, "grad_norm": 0.11966105550527573, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34262232.0, "reward": 0.6565508842468262, "reward_std": 0.45658430457115173, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2505999803543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8362499475479126, "step": 2919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 410.9375, "completions/mean_terminated_length": 410.9375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.584, "grad_norm": 5.4655192798236385e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34273479.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 498.6875, "completions/mean_terminated_length": 498.6875, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.5842, "grad_norm": 0.00999381858855486, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34287346.0, "reward": 0.5572665333747864, "reward_std": 0.02460188791155815, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1145312488079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "step": 2921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 331.3125, "completions/mean_terminated_length": 331.3125, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.5844, "grad_norm": 0.00016020869952626526, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34296295.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 712.375, "completions/mean_terminated_length": 667.857177734375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5846, "grad_norm": 0.12440742552280426, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34311997.0, "reward": 0.8153911828994751, "reward_std": 0.5106562376022339, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5682812333106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4906249940395355, "step": 2923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 723.25, "completions/mean_terminated_length": 703.2000122070312, "completions/min_length": 611.0, "completions/min_terminated_length": 611.0, "epoch": 0.5848, "grad_norm": 0.12051814049482346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34327873.0, "reward": 0.9848040342330933, "reward_std": 0.5470123887062073, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5321062803268433, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 2924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 381.3125, "completions/mean_terminated_length": 381.3125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.585, "grad_norm": 0.15641513466835022, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34338070.0, "reward": 1.1959260702133179, "reward_std": 0.4616238474845886, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7043499946594238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 2925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 419.4375, "completions/mean_terminated_length": 419.4375, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.5852, "grad_norm": 0.012943173758685589, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34348941.0, "reward": 0.5447636842727661, "reward_std": 0.0514177642762661, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0895249992609024, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.95250004529953, "step": 2926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 283.25, "completions/mean_terminated_length": 283.25, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.5854, "grad_norm": 0.07283459603786469, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34357665.0, "reward": 1.4644384384155273, "reward_std": 0.12548120319843292, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9913749694824219, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8825000524520874, "step": 2927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 303.1875, "completions/mean_terminated_length": 303.1875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.5856, "grad_norm": 0.3240112066268921, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34367012.0, "reward": 1.2241791486740112, "reward_std": 0.5061304569244385, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7608562707901001, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 788.875, "completions/mean_terminated_length": 647.7999877929688, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.5858, "grad_norm": 0.17458869516849518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34384754.0, "reward": 0.7163287401199341, "reward_std": 0.6769614815711975, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4326562285423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.565625011920929, "step": 2929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 396.375, "completions/mean_terminated_length": 396.375, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.586, "grad_norm": 0.14217621088027954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34395720.0, "reward": 0.9887383580207825, "reward_std": 0.44001567363739014, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6024750471115112, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7862499952316284, "step": 2930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 503.625, "completions/mean_terminated_length": 503.625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.5862, "grad_norm": 0.13922075927257538, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34407826.0, "reward": 0.7832041382789612, "reward_std": 0.42683982849121094, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 447.75, "completions/mean_terminated_length": 447.75, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.5864, "grad_norm": 0.1670059710741043, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34418574.0, "reward": 0.9813915491104126, "reward_std": 0.4726220965385437, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5252812504768372, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9356250166893005, "step": 2932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 600.9375, "completions/mean_terminated_length": 572.7333374023438, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.5866, "grad_norm": 0.07850395888090134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34432589.0, "reward": 0.7556256055831909, "reward_std": 0.29225826263427734, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5737500190734863, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 2933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 977.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 691.1875, "completions/mean_terminated_length": 691.1875, "completions/min_length": 423.0, "completions/min_terminated_length": 423.0, "epoch": 0.5868, "grad_norm": 0.0726906955242157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34447760.0, "reward": 0.9908351302146912, "reward_std": 0.3651505410671234, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6691687703132629, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6131249666213989, "step": 2934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 573.1875, "completions/mean_terminated_length": 573.1875, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.587, "grad_norm": 0.06295480579137802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34460883.0, "reward": 0.9053909778594971, "reward_std": 0.1229468509554863, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8107812404632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.359375, "step": 2935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 502.0, "completions/max_terminated_length": 502.0, "completions/mean_length": 335.375, "completions/mean_terminated_length": 335.375, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.5872, "grad_norm": 0.17270950973033905, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34469961.0, "reward": 0.7779543399810791, "reward_std": 0.4306943416595459, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3059062361717224, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 2936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 346.125, "completions/mean_terminated_length": 346.125, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.5874, "grad_norm": 0.12734699249267578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34479851.0, "reward": 1.3800791501998901, "reward_std": 0.3245174288749695, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851562738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 253.0625, "completions/mean_terminated_length": 253.0625, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.5876, "grad_norm": 0.14273546636104584, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34487532.0, "reward": 1.388485312461853, "reward_std": 0.30313044786453247, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.901968777179718, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 2938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 562.75, "completions/mean_terminated_length": 562.75, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5878, "grad_norm": 0.00524657778441906, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34500856.0, "reward": 0.5527228116989136, "reward_std": 0.0180070698261261, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10544374585151672, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 2939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 741.5, "completions/mean_terminated_length": 741.5, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.588, "grad_norm": 0.0902523547410965, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34517104.0, "reward": 1.014272928237915, "reward_std": 0.466156542301178, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5910437703132629, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8568750023841858, "step": 2940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 200.625, "completions/mean_terminated_length": 200.625, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.5882, "grad_norm": 0.562809944152832, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34524074.0, "reward": 1.2898105382919312, "reward_std": 0.46669235825538635, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8296187520027161, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9706249833106995, "step": 2941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 561.75, "completions/mean_terminated_length": 530.933349609375, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.5884, "grad_norm": 0.19328844547271729, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34536726.0, "reward": 0.7948412895202637, "reward_std": 0.36843258142471313, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5271812677383423, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6118749976158142, "step": 2942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 195.0, "completions/max_terminated_length": 195.0, "completions/mean_length": 151.1875, "completions/mean_terminated_length": 151.1875, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.5886, "grad_norm": 0.18910953402519226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34542937.0, "reward": 0.5820479989051819, "reward_std": 0.245851531624794, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10159356892108917, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9766812324523926, "step": 2943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 324.6875, "completions/mean_terminated_length": 324.6875, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.5888, "grad_norm": 0.1304139643907547, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34552100.0, "reward": 1.439450979232788, "reward_std": 0.23752017319202423, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9414000511169434, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 2944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 281.3125, "completions/mean_terminated_length": 281.3125, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.589, "grad_norm": 0.10014217346906662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34560873.0, "reward": 1.4395290613174438, "reward_std": 0.23754122853279114, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415562152862549, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 2945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 442.0625, "completions/mean_terminated_length": 442.0625, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.5892, "grad_norm": 0.15794602036476135, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34572522.0, "reward": 1.0730476379394531, "reward_std": 0.45220544934272766, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6460937261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 2946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 416.1875, "completions/mean_terminated_length": 416.1875, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.5894, "grad_norm": 0.06947660446166992, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34582893.0, "reward": 1.4396510124206543, "reward_std": 0.23757384717464447, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9417999982833862, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 2947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 440.4375, "completions/mean_terminated_length": 440.4375, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.5896, "grad_norm": 0.21482199430465698, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34594068.0, "reward": 0.7847542762756348, "reward_std": 0.428280234336853, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3195062279701233, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 2948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 349.375, "completions/mean_terminated_length": 349.375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.5898, "grad_norm": 0.12968096137046814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34603338.0, "reward": 0.7268759608268738, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26624998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 2949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 424.1875, "completions/mean_terminated_length": 424.1875, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.59, "grad_norm": 0.08372066169977188, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34614141.0, "reward": 1.437851071357727, "reward_std": 0.24481387436389923, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9381999969482422, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 597.6875, "completions/mean_terminated_length": 597.6875, "completions/min_length": 488.0, "completions/min_terminated_length": 488.0, "epoch": 0.5902, "grad_norm": 0.1195327639579773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34627672.0, "reward": 1.1479946374893188, "reward_std": 0.39476069808006287, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7959874868392944, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7100000381469727, "step": 2951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 368.4375, "completions/mean_terminated_length": 368.4375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.5904, "grad_norm": 0.09508469700813293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34637151.0, "reward": 1.3770854473114014, "reward_std": 0.33298259973526, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8791687488555908, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 2952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 699.875, "completions/mean_terminated_length": 678.2667236328125, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.5906, "grad_norm": 0.09252840280532837, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34653117.0, "reward": 0.5964852571487427, "reward_std": 0.22726131975650787, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2554687261581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.765625, "step": 2953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 900.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 527.5625, "completions/mean_terminated_length": 527.5625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5908, "grad_norm": 0.0028375587426126003, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34666950.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 517.375, "completions/mean_terminated_length": 517.375, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.591, "grad_norm": 0.06601259857416153, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34680476.0, "reward": 0.6045071482658386, "reward_std": 0.23866795003414154, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1465124934911728, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 2955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 665.3125, "completions/mean_terminated_length": 641.4000244140625, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.5912, "grad_norm": 0.03867017477750778, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34695313.0, "reward": 0.6347508430480957, "reward_std": 0.29061025381088257, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2695000171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8199999928474426, "step": 2956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 428.1875, "completions/mean_terminated_length": 428.1875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.5914, "grad_norm": 0.12718510627746582, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34706212.0, "reward": 1.3188416957855225, "reward_std": 0.38857516646385193, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.825181245803833, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 2957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 422.125, "completions/mean_terminated_length": 382.0000305175781, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.5916, "grad_norm": 0.19253528118133545, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34716838.0, "reward": 1.0717976093292236, "reward_std": 0.49819982051849365, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7060937285423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.734375, "step": 2958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 484.1875, "completions/mean_terminated_length": 484.1875, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.5918, "grad_norm": 0.10403931885957718, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34728521.0, "reward": 1.3777544498443604, "reward_std": 0.33324381709098816, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8805062770843506, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 2959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 944.9375, "completions/mean_terminated_length": 897.5, "completions/min_length": 756.0, "completions/min_terminated_length": 756.0, "epoch": 0.592, "grad_norm": 0.1388091742992401, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34747832.0, "reward": 0.7479536533355713, "reward_std": 0.6782920360565186, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.495906263589859, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5106250047683716, "step": 2960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 946.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 677.75, "completions/mean_terminated_length": 677.75, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.5922, "grad_norm": 0.06274029612541199, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34763412.0, "reward": 0.8605473637580872, "reward_std": 0.189009428024292, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7210937738418579, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.40937501192092896, "step": 2961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 749.0, "completions/mean_terminated_length": 709.7142944335938, "completions/min_length": 561.0, "completions/min_terminated_length": 561.0, "epoch": 0.5924, "grad_norm": 0.04481993243098259, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34779460.0, "reward": 0.4806135296821594, "reward_std": 0.19119229912757874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0862250030040741, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8299999833106995, "step": 2962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 728.5, "completions/mean_terminated_length": 686.2857666015625, "completions/min_length": 500.0, "completions/min_terminated_length": 500.0, "epoch": 0.5926, "grad_norm": 0.1866803616285324, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34795196.0, "reward": 0.8801538944244385, "reward_std": 0.5480940937995911, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5103062391281128, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7618750333786011, "step": 2963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 495.375, "completions/mean_terminated_length": 495.375, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.5928, "grad_norm": 0.12203265726566315, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34806690.0, "reward": 1.0833321809768677, "reward_std": 0.486918181180954, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6041625142097473, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 2964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 413.9375, "completions/mean_terminated_length": 413.9375, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.593, "grad_norm": 0.14203065633773804, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34817281.0, "reward": 1.1707041263580322, "reward_std": 0.45014187693595886, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7164062857627869, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 2965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 864.4375, "completions/mean_terminated_length": 791.9091186523438, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.5932, "grad_norm": 0.1414719820022583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34835240.0, "reward": 0.5618601441383362, "reward_std": 0.5307268500328064, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2487187385559082, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6543749570846558, "step": 2966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 634.5, "completions/mean_terminated_length": 578.857177734375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.5934, "grad_norm": 0.17164959013462067, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34849760.0, "reward": 0.9999226331710815, "reward_std": 0.5652196407318115, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6248437762260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.746874988079071, "step": 2967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 768.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 519.625, "completions/mean_terminated_length": 519.625, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.5936, "grad_norm": 0.013679399155080318, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34862330.0, "reward": 0.5598321557044983, "reward_std": 0.05254356563091278, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11966249346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 2968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 493.9375, "completions/mean_terminated_length": 493.9375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.5938, "grad_norm": 0.15077002346515656, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34874441.0, "reward": 1.0207040309906006, "reward_std": 0.4938649833202362, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5414062738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 979.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 409.375, "completions/mean_terminated_length": 409.375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.594, "grad_norm": 0.19947148859500885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34885039.0, "reward": 1.0435917377471924, "reward_std": 0.42769765853881836, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6496812105178833, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7618750333786011, "step": 2970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 399.0, "completions/max_terminated_length": 399.0, "completions/mean_length": 332.875, "completions/mean_terminated_length": 332.875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.5942, "grad_norm": 8.898638770915568e-05, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34894557.0, "reward": 1.4988291263580322, "reward_std": 0.00031275750370696187, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976562261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 413.5, "completions/mean_terminated_length": 413.5, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.5944, "grad_norm": 0.1908728927373886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34905349.0, "reward": 1.3262510299682617, "reward_std": 0.37134891748428345, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8400000333786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 2972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 711.9375, "completions/mean_terminated_length": 691.1333618164062, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.5946, "grad_norm": 0.1016554981470108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34920740.0, "reward": 1.1833102703094482, "reward_std": 0.5100078582763672, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7416187524795532, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8731249570846558, "step": 2973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 564.0, "completions/mean_terminated_length": 564.0, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.5948, "grad_norm": 0.08232792466878891, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34934164.0, "reward": 0.6752319931983948, "reward_std": 0.2083129733800888, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.35046249628067017, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7150000333786011, "step": 2974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 382.0, "completions/max_terminated_length": 382.0, "completions/mean_length": 314.625, "completions/mean_terminated_length": 314.625, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.595, "grad_norm": 0.22872942686080933, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34943150.0, "reward": 0.9019541144371033, "reward_std": 0.4775867164134979, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42890626192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 504.125, "completions/mean_terminated_length": 504.125, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.5952, "grad_norm": 0.12622864544391632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34955216.0, "reward": 0.9155261516571045, "reward_std": 0.42339202761650085, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.518549919128418, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8736875057220459, "step": 2976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 327.0, "completions/mean_terminated_length": 327.0, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.5954, "grad_norm": 0.00023461644013877958, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34964256.0, "reward": 1.4989571571350098, "reward_std": 0.0004445721278898418, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979125261306763, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 2977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 501.125, "completions/mean_terminated_length": 501.125, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.5956, "grad_norm": 0.07755240052938461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34977666.0, "reward": 0.6937508583068848, "reward_std": 0.19293518364429474, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.38749998807907104, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7249999642372131, "step": 2978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 248.5625, "completions/mean_terminated_length": 248.5625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.5958, "grad_norm": 0.14152513444423676, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34985035.0, "reward": 1.3802947998046875, "reward_std": 0.3246018886566162, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855875134468079, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 2979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 395.875, "completions/mean_terminated_length": 395.875, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.596, "grad_norm": 0.14609310030937195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34995433.0, "reward": 1.278491735458374, "reward_std": 0.3984340727329254, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8069812655448914, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.893125057220459, "step": 2980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 960.8125, "completions/mean_terminated_length": 855.5, "completions/min_length": 679.0, "completions/min_terminated_length": 679.0, "epoch": 0.5962, "grad_norm": 0.18415507674217224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35015270.0, "reward": 0.2964814603328705, "reward_std": 0.4153099060058594, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21796250343322754, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.20250000059604645, "step": 2981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 595.25, "completions/mean_terminated_length": 566.6666870117188, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.5964, "grad_norm": 0.11210913211107254, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35028874.0, "reward": 0.9844446182250977, "reward_std": 0.5099308490753174, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.593887448310852, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8137500286102295, "step": 2982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 742.875, "completions/mean_terminated_length": 702.7142944335938, "completions/min_length": 507.0, "completions/min_terminated_length": 507.0, "epoch": 0.5966, "grad_norm": 0.06491068750619888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35045000.0, "reward": 0.7162505984306335, "reward_std": 0.3659537732601166, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4949999749660492, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5687500238418579, "step": 2983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 675.5, "completions/mean_terminated_length": 675.5, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.5968, "grad_norm": 0.11621861159801483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35059712.0, "reward": 0.8999758958816528, "reward_std": 0.4430466890335083, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4874500036239624, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8412500023841858, "step": 2984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 361.75, "completions/mean_terminated_length": 361.75, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.597, "grad_norm": 0.20290234684944153, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35071116.0, "reward": 0.8856825232505798, "reward_std": 0.49105533957481384, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3963624835014343, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 2985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 509.3125, "completions/mean_terminated_length": 509.3125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.5972, "grad_norm": 0.010033480823040009, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35082945.0, "reward": 0.5378011465072632, "reward_std": 0.0200816597789526, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07559999823570251, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 2986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 648.5625, "completions/mean_terminated_length": 623.5333862304688, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.5974, "grad_norm": 0.09948300570249557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35098730.0, "reward": 0.7559350728988647, "reward_std": 0.297593891620636, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5743687152862549, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.46812501549720764, "step": 2987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 440.875, "completions/mean_terminated_length": 440.875, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.5976, "grad_norm": 0.19141846895217896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35109768.0, "reward": 1.000378966331482, "reward_std": 0.42495808005332947, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6257562637329102, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7893750071525574, "step": 2988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 602.5, "completions/mean_terminated_length": 602.5, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.5978, "grad_norm": 0.046689752489328384, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35123264.0, "reward": 0.8261724710464478, "reward_std": 0.15104970335960388, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.65234375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5218750238418579, "step": 2989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 566.0625, "completions/mean_terminated_length": 566.0625, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.598, "grad_norm": 0.14300274848937988, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35136289.0, "reward": 1.103060245513916, "reward_std": 0.4708065390586853, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6436187624931335, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 2990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 364.0, "completions/mean_length": 348.6875, "completions/mean_terminated_length": 303.66668701171875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.5982, "grad_norm": 0.19326108694076538, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35145612.0, "reward": 1.225210189819336, "reward_std": 0.504143238067627, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7629187703132629, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 2991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 786.9375, "completions/mean_terminated_length": 771.1333618164062, "completions/min_length": 578.0, "completions/min_terminated_length": 578.0, "epoch": 0.5984, "grad_norm": 0.03655310720205307, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35163131.0, "reward": 0.5237383842468262, "reward_std": 0.14406508207321167, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10997499525547028, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8799999952316284, "step": 2992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 408.0, "completions/mean_terminated_length": 408.0, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.5986, "grad_norm": 0.18573404848575592, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35173947.0, "reward": 1.3065601587295532, "reward_std": 0.34784621000289917, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8631187677383423, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8556249737739563, "step": 2993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 511.1875, "completions/mean_terminated_length": 511.1875, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.5988, "grad_norm": 0.1851348727941513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35186222.0, "reward": 1.3830227851867676, "reward_std": 0.3167520761489868, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8910437226295471, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 2994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 366.1875, "completions/mean_terminated_length": 366.1875, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.599, "grad_norm": 0.05445422977209091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35196465.0, "reward": 1.4394540786743164, "reward_std": 0.23752105236053467, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.94140625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 2995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 519.0625, "completions/mean_terminated_length": 519.0625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.5992, "grad_norm": 0.11311092227697372, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35209250.0, "reward": 0.8128132820129395, "reward_std": 0.21596723794937134, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6256250143051147, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.44999998807907104, "step": 2996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 821.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 591.8125, "completions/mean_terminated_length": 591.8125, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.5994, "grad_norm": 0.14476285874843597, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35222719.0, "reward": 1.184297800064087, "reward_std": 0.4319973289966583, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.743593692779541, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 2997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 370.5625, "completions/mean_terminated_length": 370.5625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.5996, "grad_norm": 0.02150433324277401, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35232840.0, "reward": 0.9675002098083496, "reward_std": 0.04905766248703003, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9350000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.21250000596046448, "step": 2998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 421.4375, "completions/mean_terminated_length": 421.4375, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.5998, "grad_norm": 0.14787372946739197, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35244031.0, "reward": 0.7768763303756714, "reward_std": 0.4310167133808136, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3037499785423279, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 2999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 395.3125, "completions/mean_terminated_length": 395.3125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.6, "grad_norm": 0.04453904554247856, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35254660.0, "reward": 0.61500084400177, "reward_std": 0.10323706269264221, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22999998927116394, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 3000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 981.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 734.0, "completions/mean_terminated_length": 734.0, "completions/min_length": 460.0, "completions/min_terminated_length": 460.0, "epoch": 0.6002, "grad_norm": 0.053775496780872345, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35270404.0, "reward": 0.7360162734985352, "reward_std": 0.17711089551448822, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47203123569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6656249761581421, "step": 3001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 325.5, "completions/mean_terminated_length": 325.5, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.6004, "grad_norm": 0.21950098872184753, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35280044.0, "reward": 1.2865386009216309, "reward_std": 0.47121894359588623, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8230750560760498, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 3002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 702.0, "completions/mean_terminated_length": 680.5333862304688, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.6006, "grad_norm": 0.08042217791080475, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35295292.0, "reward": 1.3449195623397827, "reward_std": 0.429747611284256, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8773374557495117, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8837499618530273, "step": 3003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 349.5625, "completions/mean_terminated_length": 349.5625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.6008, "grad_norm": 0.16001226007938385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35305413.0, "reward": 1.3801414966583252, "reward_std": 0.3245418965816498, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8852812647819519, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 471.75, "completions/mean_terminated_length": 471.75, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.601, "grad_norm": 0.21929147839546204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35317185.0, "reward": 1.1058601140975952, "reward_std": 0.4293126165866852, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.711718738079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 3005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 645.875, "completions/mean_terminated_length": 620.6666870117188, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.6012, "grad_norm": 0.054970379918813705, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35331807.0, "reward": 0.7120318412780762, "reward_std": 0.24617648124694824, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4865625202655792, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.59375, "step": 3006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 639.6875, "completions/mean_terminated_length": 584.7857666015625, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.6014, "grad_norm": 0.13267384469509125, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35346554.0, "reward": 0.6507036685943604, "reward_std": 0.3303939700126648, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.42640623450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.503125011920929, "step": 3007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 208.0, "completions/max_terminated_length": 208.0, "completions/mean_length": 176.5625, "completions/mean_terminated_length": 176.5625, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.6016, "grad_norm": 0.00044822381460107863, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35353123.0, "reward": 1.4994980096817017, "reward_std": 0.0005991500802338123, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989937543869019, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 3008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 407.625, "completions/mean_terminated_length": 407.625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.6018, "grad_norm": 0.09172835946083069, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35363005.0, "reward": 0.7195320129394531, "reward_std": 0.17450137436389923, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4390624761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 3009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 439.375, "completions/mean_terminated_length": 439.375, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.602, "grad_norm": 0.12749610841274261, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35374339.0, "reward": 1.2268635034561157, "reward_std": 0.5013983249664307, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7662249803543091, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 436.0, "completions/mean_terminated_length": 436.0, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.6022, "grad_norm": 0.09864213317632675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35385651.0, "reward": 0.6644541621208191, "reward_std": 0.32590189576148987, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20390623807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 301.3125, "completions/mean_terminated_length": 301.3125, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.6024, "grad_norm": 0.14617836475372314, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35394408.0, "reward": 1.3233604431152344, "reward_std": 0.3767926096916199, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8342187404632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 479.625, "completions/mean_terminated_length": 479.625, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.6026, "grad_norm": 0.30139273405075073, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35405538.0, "reward": 0.7249821424484253, "reward_std": 0.38451066613197327, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2624624967575073, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 382.3125, "completions/mean_terminated_length": 382.3125, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.6028, "grad_norm": 0.1302277147769928, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35415607.0, "reward": 1.2507703304290771, "reward_std": 0.4447426199913025, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7515374422073364, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 3014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 754.75, "completions/mean_terminated_length": 665.0, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.603, "grad_norm": 0.1596779227256775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35432771.0, "reward": 0.6255477666854858, "reward_std": 0.5193213820457458, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3135937452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6531250476837158, "step": 3015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 756.75, "completions/mean_terminated_length": 667.6666870117188, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.6032, "grad_norm": 0.12207047641277313, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35449279.0, "reward": 0.7406256198883057, "reward_std": 0.5865186452865601, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41874998807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.65625, "step": 3016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 442.4375, "completions/mean_terminated_length": 442.4375, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6034, "grad_norm": 0.0762770026922226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35460774.0, "reward": 0.5699759721755981, "reward_std": 0.1143045425415039, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1399500072002411, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903749942779541, "step": 3017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 573.125, "completions/mean_terminated_length": 573.125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.6036, "grad_norm": 0.17066386342048645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35474600.0, "reward": 1.25217604637146, "reward_std": 0.4413776397705078, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7543500065803528, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 3018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 229.8125, "completions/mean_terminated_length": 229.8125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.6038, "grad_norm": 0.08648376166820526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35483525.0, "reward": 1.439804196357727, "reward_std": 0.23761487007141113, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421062469482422, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 357.125, "completions/mean_terminated_length": 357.125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.604, "grad_norm": 0.07026708126068115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35493543.0, "reward": 0.5990415811538696, "reward_std": 0.24047008156776428, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13558125495910645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 373.75, "completions/mean_terminated_length": 330.4000244140625, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 0.6042, "grad_norm": 0.043656863272190094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35503875.0, "reward": 0.5288258790969849, "reward_std": 0.15798160433769226, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12014999240636826, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8712500333786011, "step": 3021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 676.625, "completions/mean_terminated_length": 676.625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.6044, "grad_norm": 0.08067045360803604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35518717.0, "reward": 0.7486728429794312, "reward_std": 0.3360886573791504, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3723437786102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.809374988079071, "step": 3022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 440.25, "completions/mean_terminated_length": 401.3333435058594, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.6046, "grad_norm": 0.1072484478354454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35530881.0, "reward": 0.877109706401825, "reward_std": 0.2743073105812073, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8167188167572021, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.203125, "step": 3023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 819.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 487.0625, "completions/mean_terminated_length": 487.0625, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.6048, "grad_norm": 0.07610166817903519, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35542690.0, "reward": 0.6470321416854858, "reward_std": 0.2508692145347595, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23156249523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 3024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 524.9375, "completions/mean_terminated_length": 524.9375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.605, "grad_norm": 0.1824634075164795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35554897.0, "reward": 0.7560166120529175, "reward_std": 0.3773053288459778, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3245312571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 3025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 305.9375, "completions/mean_terminated_length": 305.9375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.6052, "grad_norm": 0.12843096256256104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35563904.0, "reward": 1.4372260570526123, "reward_std": 0.24727384746074677, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9369500279426575, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 3026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 747.5625, "completions/mean_terminated_length": 683.7692260742188, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.6054, "grad_norm": 0.09826180338859558, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35579801.0, "reward": 0.4420320391654968, "reward_std": 0.28657764196395874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1340624988079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.65625, "step": 3027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 498.25, "completions/mean_terminated_length": 498.25, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.6056, "grad_norm": 0.05433989688754082, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35591517.0, "reward": 1.4597665071487427, "reward_std": 0.15593762695789337, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9820312261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 3028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 819.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 667.8125, "completions/mean_terminated_length": 667.8125, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.6058, "grad_norm": 0.09745433181524277, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35606506.0, "reward": 0.9936724901199341, "reward_std": 0.2940908372402191, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7998437881469727, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4781250059604645, "step": 3029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 972.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 651.0, "completions/mean_terminated_length": 651.0, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.606, "grad_norm": 0.055509183555841446, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35620986.0, "reward": 0.7036726474761963, "reward_std": 0.16888800263404846, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4073437452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 3030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 395.5625, "completions/mean_terminated_length": 395.5625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.6062, "grad_norm": 0.2467648684978485, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35632099.0, "reward": 0.9000664353370667, "reward_std": 0.4324556589126587, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42513126134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8206250071525574, "step": 3031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 432.1875, "completions/mean_terminated_length": 432.1875, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.6064, "grad_norm": 0.0983133614063263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35643126.0, "reward": 1.2613165378570557, "reward_std": 0.4248921275138855, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7726312875747681, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 345.125, "completions/mean_terminated_length": 345.125, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.6066, "grad_norm": 0.29399970173835754, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35653176.0, "reward": 1.0237510204315186, "reward_std": 0.49057790637016296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 293.4375, "completions/mean_terminated_length": 293.4375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.6068, "grad_norm": 0.17703497409820557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35661615.0, "reward": 1.1334667205810547, "reward_std": 0.48860299587249756, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6419312953948975, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831250309944153, "step": 3034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 448.9375, "completions/mean_terminated_length": 448.9375, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.607, "grad_norm": 0.18398478627204895, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35672222.0, "reward": 1.2048321962356567, "reward_std": 0.45052218437194824, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7221624851226807, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 384.75, "completions/mean_terminated_length": 384.75, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.6072, "grad_norm": 0.16144931316375732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35682218.0, "reward": 0.6675009727478027, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20999999344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 890.4375, "completions/mean_terminated_length": 786.5555419921875, "completions/min_length": 523.0, "completions/min_terminated_length": 523.0, "epoch": 0.6074, "grad_norm": 0.09894904494285583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35700817.0, "reward": 0.36765676736831665, "reward_std": 0.3517788350582123, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17281249165534973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 3037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 372.4375, "completions/mean_terminated_length": 372.4375, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.6076, "grad_norm": 0.17276890575885773, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35710312.0, "reward": 0.964375913143158, "reward_std": 0.4867301285266876, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49125000834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 388.0, "completions/max_terminated_length": 388.0, "completions/mean_length": 307.25, "completions/mean_terminated_length": 307.25, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.6078, "grad_norm": 0.18711963295936584, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35719052.0, "reward": 0.7914854288101196, "reward_std": 0.4221751391887665, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33296874165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "step": 3039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 971.75, "completions/mean_terminated_length": 919.5, "completions/min_length": 731.0, "completions/min_terminated_length": 731.0, "epoch": 0.608, "grad_norm": 0.18668785691261292, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35739032.0, "reward": 0.48648470640182495, "reward_std": 0.5636275410652161, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3479687571525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 3040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 380.1875, "completions/mean_terminated_length": 380.1875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6082, "grad_norm": 0.00021346728317439556, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35748939.0, "reward": 1.4990510940551758, "reward_std": 0.0005388545687310398, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980999827384949, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 3041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 363.625, "completions/mean_terminated_length": 363.625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.6084, "grad_norm": 0.20157326757907867, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35758661.0, "reward": 1.0860165357589722, "reward_std": 0.48347142338752747, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6095312833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 844.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 598.0, "completions/mean_terminated_length": 598.0, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.6086, "grad_norm": 0.052190229296684265, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35772549.0, "reward": 0.8392850160598755, "reward_std": 0.19252732396125793, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6785687208175659, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.43312501907348633, "step": 3043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 348.5625, "completions/mean_terminated_length": 348.5625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.6088, "grad_norm": 0.20450006425380707, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35782142.0, "reward": 0.9614854454994202, "reward_std": 0.48983234167099, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4854687452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 3044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 305.3125, "completions/mean_terminated_length": 305.3125, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.609, "grad_norm": 0.24759656190872192, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35791155.0, "reward": 0.7855227589607239, "reward_std": 0.4255851209163666, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3210437595844269, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 3045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 529.9375, "completions/mean_terminated_length": 497.0000305175781, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.6092, "grad_norm": 0.07087286561727524, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35803970.0, "reward": 0.5529695749282837, "reward_std": 0.16800305247306824, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16843748092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "step": 3046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 470.0, "completions/mean_terminated_length": 470.0, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.6094, "grad_norm": 0.14730197191238403, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35815746.0, "reward": 0.791472852230072, "reward_std": 0.423397034406662, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33294373750686646, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456249475479126, "step": 3047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 655.875, "completions/mean_terminated_length": 631.3333740234375, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.6096, "grad_norm": 0.23655886948108673, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35830048.0, "reward": 0.7392975687980652, "reward_std": 0.2605864107608795, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.41609376668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.703125, "step": 3048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 914.75, "completions/mean_terminated_length": 829.7777709960938, "completions/min_length": 557.0, "completions/min_terminated_length": 557.0, "epoch": 0.6098, "grad_norm": 0.18280740082263947, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35848668.0, "reward": 0.4567973017692566, "reward_std": 0.47086432576179504, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.28859376907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.40312498807907104, "step": 3049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 348.625, "completions/mean_terminated_length": 348.625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.61, "grad_norm": 0.2511961758136749, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35858294.0, "reward": 1.0716854333877563, "reward_std": 0.5004316568374634, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5808687806129456, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 705.3125, "completions/mean_terminated_length": 631.7692260742188, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.6102, "grad_norm": 0.10540185123682022, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35873275.0, "reward": 0.6360099911689758, "reward_std": 0.44711366295814514, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.33451876044273376, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6643750071525574, "step": 3051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 433.875, "completions/mean_terminated_length": 433.875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.6104, "grad_norm": 0.15383009612560272, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35884601.0, "reward": 1.3410165309906006, "reward_std": 0.3460233211517334, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8695312738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 3052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 871.0, "completions/mean_terminated_length": 779.2000122070312, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.6106, "grad_norm": 0.11484795808792114, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35903209.0, "reward": 0.5242192149162292, "reward_std": 0.48448628187179565, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3609375059604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.36250001192092896, "step": 3053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 279.0, "completions/max_terminated_length": 279.0, "completions/mean_length": 228.0, "completions/mean_terminated_length": 228.0, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.6108, "grad_norm": 0.15337355434894562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35910553.0, "reward": 1.3798447847366333, "reward_std": 0.3244273364543915, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8846874833106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 409.875, "completions/mean_terminated_length": 409.875, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.611, "grad_norm": 0.0847373828291893, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35921271.0, "reward": 0.6026541590690613, "reward_std": 0.23942889273166656, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428062468767166, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 557.0, "completions/mean_terminated_length": 557.0, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.6112, "grad_norm": 0.05852947384119034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35933815.0, "reward": 0.7321882247924805, "reward_std": 0.12884362041950226, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46437501907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.706250011920929, "step": 3056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 578.125, "completions/mean_terminated_length": 548.4000244140625, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.6114, "grad_norm": 0.1421845406293869, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35946809.0, "reward": 0.7530475854873657, "reward_std": 0.46351850032806396, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.44359374046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.731249988079071, "step": 3057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 442.75, "completions/mean_terminated_length": 442.75, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.6116, "grad_norm": 0.17556986212730408, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35958293.0, "reward": 0.7268759608268738, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26624998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 519.375, "completions/mean_terminated_length": 447.2857360839844, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.6118, "grad_norm": 0.1802789717912674, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35970971.0, "reward": 1.0400009155273438, "reward_std": 0.5791343450546265, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6425000429153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.793749988079071, "step": 3059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 631.8125, "completions/mean_terminated_length": 575.7857666015625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.612, "grad_norm": 0.09688995778560638, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35984760.0, "reward": 0.6527351140975952, "reward_std": 0.31592613458633423, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.43046873807907104, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5218750238418579, "step": 3060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 261.5625, "completions/mean_terminated_length": 261.5625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.6122, "grad_norm": 0.22664573788642883, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35993025.0, "reward": 1.3207571506500244, "reward_std": 0.3830225169658661, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290125131607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 323.375, "completions/mean_terminated_length": 323.375, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.6124, "grad_norm": 0.14392904937267303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36002487.0, "reward": 1.0832664966583252, "reward_std": 0.48685839772224426, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6040312647819519, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 677.375, "completions/mean_terminated_length": 677.375, "completions/min_length": 562.0, "completions/min_terminated_length": 562.0, "epoch": 0.6126, "grad_norm": 0.08075974136590958, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36017693.0, "reward": 1.4400447607040405, "reward_std": 0.23767909407615662, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9425874948501587, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 3063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 355.5, "completions/mean_terminated_length": 355.5, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.6128, "grad_norm": 0.1571638286113739, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36027637.0, "reward": 0.8137509822845459, "reward_std": 0.4229218065738678, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3774999976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 3064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 406.6875, "completions/mean_terminated_length": 406.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.613, "grad_norm": 0.11993865668773651, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36037744.0, "reward": 1.198828935623169, "reward_std": 0.41241130232810974, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.772656261920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 3065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 409.3125, "completions/mean_terminated_length": 409.3125, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.6132, "grad_norm": 0.2514107823371887, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36048325.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 389.0, "completions/mean_terminated_length": 389.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.6134, "grad_norm": 0.24634969234466553, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36058661.0, "reward": 1.3115633726119995, "reward_std": 0.3498314619064331, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8731250166893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.862500011920929, "step": 3067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 352.375, "completions/mean_terminated_length": 352.375, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.6136, "grad_norm": 0.1616099327802658, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36068155.0, "reward": 1.383044719696045, "reward_std": 0.31676071882247925, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.891087532043457, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 265.9375, "completions/mean_terminated_length": 265.9375, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.6138, "grad_norm": 0.18195059895515442, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36076042.0, "reward": 1.3208508491516113, "reward_std": 0.3830692172050476, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829200029373169, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 3069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 616.8125, "completions/mean_terminated_length": 616.8125, "completions/min_length": 449.0, "completions/min_terminated_length": 449.0, "epoch": 0.614, "grad_norm": 0.0704021230340004, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36092695.0, "reward": 0.8812510371208191, "reward_std": 0.44594594836235046, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.44999998807907104, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812499642372131, "step": 3070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 367.75, "completions/mean_terminated_length": 367.75, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.6142, "grad_norm": 0.09790182113647461, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36102819.0, "reward": 1.4366135597229004, "reward_std": 0.2497635781764984, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9357249736785889, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 3071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 292.5625, "completions/mean_terminated_length": 292.5625, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.6144, "grad_norm": 0.00012053851241944358, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36111532.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 635.9375, "completions/mean_terminated_length": 610.0667114257812, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.6146, "grad_norm": 0.13485409319400787, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36125963.0, "reward": 0.699504017829895, "reward_std": 0.4217984676361084, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2740062475204468, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 3073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 409.0625, "completions/mean_terminated_length": 409.0625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.6148, "grad_norm": 0.21637223660945892, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36136460.0, "reward": 1.1035946607589722, "reward_std": 0.4694458544254303, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6446875333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 3074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 720.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 437.6875, "completions/mean_terminated_length": 437.6875, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.615, "grad_norm": 0.2818463146686554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36147239.0, "reward": 0.9727323055267334, "reward_std": 0.4805014133453369, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5079624652862549, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 3075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 783.0, "completions/mean_terminated_length": 748.5714721679688, "completions/min_length": 507.0, "completions/min_terminated_length": 507.0, "epoch": 0.6152, "grad_norm": 0.09661916643381119, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36164135.0, "reward": 0.5972509384155273, "reward_std": 0.35649752616882324, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.25699999928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7200000286102295, "step": 3076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 532.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 357.6875, "completions/mean_terminated_length": 357.6875, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.6154, "grad_norm": 0.09547725319862366, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36173586.0, "reward": 1.3444229364395142, "reward_std": 0.339339017868042, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8763437271118164, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9368749856948853, "step": 3077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 934.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 735.8125, "completions/mean_terminated_length": 735.8125, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.6156, "grad_norm": 0.03174808248877525, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36189471.0, "reward": 0.6144540905952454, "reward_std": 0.162146657705307, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22890624403953552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 3078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 686.125, "completions/mean_terminated_length": 686.125, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.6158, "grad_norm": 0.10310177505016327, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36204593.0, "reward": 0.8130226731300354, "reward_std": 0.37846091389656067, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.43854376673698425, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8068749904632568, "step": 3079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 658.4375, "completions/mean_terminated_length": 658.4375, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.616, "grad_norm": 0.1751367598772049, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36218856.0, "reward": 1.1625008583068848, "reward_std": 0.44681230187416077, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7000000476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 3080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 940.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 628.5625, "completions/mean_terminated_length": 628.5625, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.6162, "grad_norm": 0.10431645065546036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36232945.0, "reward": 0.817266583442688, "reward_std": 0.4148070514202118, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38453125953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 3081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 597.625, "completions/mean_terminated_length": 597.625, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.6164, "grad_norm": 0.14645057916641235, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36246139.0, "reward": 0.8753132820129395, "reward_std": 0.3749431371688843, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5006250143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 3082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 471.8125, "completions/mean_terminated_length": 471.8125, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.6166, "grad_norm": 0.23655490577220917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36257592.0, "reward": 0.9300789833068848, "reward_std": 0.534917414188385, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48515626788139343, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 3083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 391.5625, "completions/mean_terminated_length": 391.5625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.6168, "grad_norm": 0.135030135512352, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36269105.0, "reward": 0.7838915586471558, "reward_std": 0.42652419209480286, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3177812397480011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 653.25, "completions/mean_terminated_length": 653.25, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.617, "grad_norm": 0.1399148553609848, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36284293.0, "reward": 0.7650790214538574, "reward_std": 0.3732115924358368, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3426562547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 3085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 433.875, "completions/mean_terminated_length": 433.875, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.6172, "grad_norm": 0.16470873355865479, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36295363.0, "reward": 0.891482412815094, "reward_std": 0.4867766797542572, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40796250104904175, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 3086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 708.0, "completions/max_terminated_length": 708.0, "completions/mean_length": 525.9375, "completions/mean_terminated_length": 525.9375, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.6174, "grad_norm": 0.21511417627334595, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36307154.0, "reward": 1.020938754081726, "reward_std": 0.49410736560821533, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5418750047683716, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 3087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 687.6875, "completions/mean_terminated_length": 687.6875, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.6176, "grad_norm": 0.13512645661830902, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36322605.0, "reward": 0.8807321786880493, "reward_std": 0.44465386867523193, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4489625096321106, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9012500047683716, "step": 3088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 330.0625, "completions/mean_terminated_length": 330.0625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.6178, "grad_norm": 0.1489124596118927, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36332254.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 494.375, "completions/mean_terminated_length": 494.375, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.618, "grad_norm": 0.10362453013658524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36343940.0, "reward": 0.9058600664138794, "reward_std": 0.2854507863521576, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6867187023162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5593750476837158, "step": 3090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 402.5, "completions/mean_terminated_length": 402.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.6182, "grad_norm": 0.1519109606742859, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36354396.0, "reward": 1.3183414936065674, "reward_std": 0.3883269727230072, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8241812586784363, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 3091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 344.5, "completions/mean_terminated_length": 344.5, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.6184, "grad_norm": 0.13720902800559998, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36363652.0, "reward": 0.7892072200775146, "reward_std": 0.4234011173248291, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3284124732017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 480.5625, "completions/mean_terminated_length": 480.5625, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.6186, "grad_norm": 0.09845738112926483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36375325.0, "reward": 0.6132822036743164, "reward_std": 0.23398710787296295, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1640625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 3093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 317.0, "completions/max_terminated_length": 317.0, "completions/mean_length": 269.875, "completions/mean_terminated_length": 269.875, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.6188, "grad_norm": 0.25924891233444214, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36383307.0, "reward": 1.3801416158676147, "reward_std": 0.3245418965816498, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8852812647819519, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 416.375, "completions/mean_terminated_length": 416.375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.619, "grad_norm": 0.2848275899887085, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36393665.0, "reward": 1.0500009059906006, "reward_std": 0.47443825006484985, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6000000238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 3095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 676.0, "completions/max_terminated_length": 676.0, "completions/mean_length": 373.0625, "completions/mean_terminated_length": 373.0625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.6192, "grad_norm": 0.17356331646442413, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36403522.0, "reward": 1.4005978107452393, "reward_std": 0.2753692865371704, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9261937737464905, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 3096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 441.75, "completions/mean_terminated_length": 441.75, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.6194, "grad_norm": 0.2132873237133026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36414798.0, "reward": 1.0860165357589722, "reward_std": 0.48347142338752747, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6095312833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 365.75, "completions/mean_terminated_length": 321.8666687011719, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.6196, "grad_norm": 0.269086092710495, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36424522.0, "reward": 0.6977323293685913, "reward_std": 0.422869473695755, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2704625129699707, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 3098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 605.5, "completions/mean_terminated_length": 605.5, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.6198, "grad_norm": 0.07522432506084442, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36438002.0, "reward": 0.5988293886184692, "reward_std": 0.24138757586479187, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13515624403953552, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 3099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 438.625, "completions/mean_terminated_length": 438.625, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.62, "grad_norm": 0.26024389266967773, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36448748.0, "reward": 0.8780351877212524, "reward_std": 0.4158582389354706, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.44356876611709595, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8456249833106995, "step": 3100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 334.875, "completions/mean_terminated_length": 288.933349609375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.6202, "grad_norm": 0.4476252496242523, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36457978.0, "reward": 0.8064727783203125, "reward_std": 0.500274121761322, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36294376850128174, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 3101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 364.125, "completions/mean_terminated_length": 364.125, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.6204, "grad_norm": 0.3020569980144501, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36467436.0, "reward": 1.2611603736877441, "reward_std": 0.4255252480506897, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7723187804222107, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 977.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 681.6875, "completions/mean_terminated_length": 681.6875, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.6206, "grad_norm": 0.10152798891067505, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36482215.0, "reward": 0.6798977851867676, "reward_std": 0.3215567469596863, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2347937524318695, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9381250143051147, "step": 3103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 556.3125, "completions/mean_terminated_length": 556.3125, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.6208, "grad_norm": 0.00866351742297411, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36496492.0, "reward": 0.5323853492736816, "reward_std": 0.019645363092422485, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06476874649524689, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 3104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1007.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 464.4375, "completions/mean_terminated_length": 464.4375, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.621, "grad_norm": 0.10062886029481888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36507891.0, "reward": 0.7719537615776062, "reward_std": 0.225070059299469, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5439062714576721, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5281250476837158, "step": 3105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 426.25, "completions/mean_terminated_length": 426.25, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.6212, "grad_norm": 0.11270584911108017, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36518743.0, "reward": 1.3168916702270508, "reward_std": 0.39237844944000244, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8212812542915344, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 3106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 779.1875, "completions/mean_terminated_length": 722.6923217773438, "completions/min_length": 528.0, "completions/min_terminated_length": 528.0, "epoch": 0.6214, "grad_norm": 0.13322676718235016, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36535962.0, "reward": 0.6434381008148193, "reward_std": 0.3728107810020447, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47437500953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4124999940395355, "step": 3107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 338.375, "completions/mean_terminated_length": 338.375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.6216, "grad_norm": 0.20569021999835968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36545328.0, "reward": 1.4004664421081543, "reward_std": 0.27531906962394714, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9259312152862549, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9243749380111694, "step": 3108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 751.3125, "completions/mean_terminated_length": 712.357177734375, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.6218, "grad_norm": 0.2070358693599701, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36561749.0, "reward": 0.8506226539611816, "reward_std": 0.4927294850349426, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5762437582015991, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5931249856948853, "step": 3109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 584.5, "completions/mean_terminated_length": 555.2000122070312, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.622, "grad_norm": 0.1602940559387207, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36574829.0, "reward": 1.1646103858947754, "reward_std": 0.5269318222999573, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7042187452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 3110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 591.625, "completions/mean_terminated_length": 562.800048828125, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.6222, "grad_norm": 0.03272676840424538, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36589063.0, "reward": 0.4979071617126465, "reward_std": 0.1362539529800415, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.058312248438596725, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9073749780654907, "step": 3111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 691.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 538.3125, "completions/mean_terminated_length": 538.3125, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.6224, "grad_norm": 0.17913572490215302, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36601308.0, "reward": 0.7975009083747864, "reward_std": 0.38214021921157837, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4074999988079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 3112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 470.0625, "completions/mean_terminated_length": 433.13336181640625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.6226, "grad_norm": 0.14132285118103027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36613437.0, "reward": 0.6864854097366333, "reward_std": 0.42503708600997925, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24796873331069946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 347.875, "completions/mean_terminated_length": 347.875, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.6228, "grad_norm": 0.003347542602568865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36622507.0, "reward": 0.5463259220123291, "reward_std": 0.009699980728328228, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09264999628067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 571.1875, "completions/mean_terminated_length": 571.1875, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.623, "grad_norm": 0.13862931728363037, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36636414.0, "reward": 0.7265634536743164, "reward_std": 0.4167376756668091, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.328125, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 3115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 582.6875, "completions/mean_terminated_length": 582.6875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.6232, "grad_norm": 0.15566949546337128, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36649561.0, "reward": 0.7146888971328735, "reward_std": 0.3903597593307495, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24187499284744263, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 3116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 786.0625, "completions/mean_terminated_length": 770.2000122070312, "completions/min_length": 566.0, "completions/min_terminated_length": 566.0, "epoch": 0.6234, "grad_norm": 0.24587300419807434, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36666650.0, "reward": 1.1634101867675781, "reward_std": 0.5283982157707214, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7018187642097473, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 795.4375, "completions/mean_terminated_length": 742.6923217773438, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.6236, "grad_norm": 0.10647915303707123, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36683297.0, "reward": 0.9816381931304932, "reward_std": 0.6110058426856995, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5882749557495117, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6712499856948853, "step": 3118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 541.0, "completions/mean_terminated_length": 541.0, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.6238, "grad_norm": 0.16936415433883667, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36696017.0, "reward": 1.1420321464538574, "reward_std": 0.4746266007423401, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6590625047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 3119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 507.0625, "completions/mean_terminated_length": 472.60003662109375, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.624, "grad_norm": 0.1726166307926178, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36708146.0, "reward": 0.9871822595596313, "reward_std": 0.5138749480247498, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5993624925613403, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 3120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 473.3125, "completions/mean_terminated_length": 473.3125, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.6242, "grad_norm": 0.011536858044564724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36719303.0, "reward": 0.5381230115890503, "reward_std": 0.03066682256758213, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07624374330043793, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606187343597412, "step": 3121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 445.5, "completions/mean_terminated_length": 445.5, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.6244, "grad_norm": 0.11091949045658112, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36730223.0, "reward": 0.704372763633728, "reward_std": 0.31990954279899597, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2837437391281128, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 3122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 515.375, "completions/mean_terminated_length": 515.375, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.6246, "grad_norm": 0.14901629090309143, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36742053.0, "reward": 0.8465384244918823, "reward_std": 0.4539841115474701, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3805750012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 3123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 373.6875, "completions/mean_terminated_length": 280.7857360839844, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.6248, "grad_norm": 0.17324601113796234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36751392.0, "reward": 1.08304762840271, "reward_std": 0.5540965795516968, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7285937666893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 482.0, "completions/mean_terminated_length": 445.86669921875, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.625, "grad_norm": 0.16434495151042938, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36762976.0, "reward": 1.0488290786743164, "reward_std": 0.5429781079292297, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.59765625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 3125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 305.6875, "completions/mean_terminated_length": 305.6875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.6252, "grad_norm": 0.24915163218975067, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36771947.0, "reward": 1.1408352851867676, "reward_std": 0.47757643461227417, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6566687822341919, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 3126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 557.9375, "completions/mean_terminated_length": 557.9375, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.6254, "grad_norm": 0.10883758217096329, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36785242.0, "reward": 0.6487509608268738, "reward_std": 0.25296810269355774, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23499998450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 3127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 460.375, "completions/mean_terminated_length": 460.375, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.6256, "grad_norm": 0.0759642943739891, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36796272.0, "reward": 0.6545290350914001, "reward_std": 0.2529296278953552, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24655625224113464, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8868750333786011, "step": 3128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 521.125, "completions/mean_terminated_length": 521.125, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.6258, "grad_norm": 0.14625224471092224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36808562.0, "reward": 0.8642728328704834, "reward_std": 0.4493631422519684, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.416043758392334, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256249666213989, "step": 3129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 275.0625, "completions/mean_terminated_length": 275.0625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.626, "grad_norm": 0.10383559763431549, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36817315.0, "reward": 1.4399197101593018, "reward_std": 0.23764562606811523, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9423375129699707, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 3130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 383.125, "completions/mean_terminated_length": 383.125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.6262, "grad_norm": 0.09695017337799072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36827381.0, "reward": 0.6081259250640869, "reward_std": 0.23749999701976776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1537500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 605.75, "completions/mean_terminated_length": 605.75, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.6264, "grad_norm": 0.025587202981114388, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36841281.0, "reward": 0.5954666137695312, "reward_std": 0.08968283236026764, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1909312605857849, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 3132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 433.375, "completions/mean_terminated_length": 433.375, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6266, "grad_norm": 0.0666809007525444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36851911.0, "reward": 0.9250003099441528, "reward_std": 0.12535429000854492, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8500000238418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.29375001788139343, "step": 3133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 698.0, "completions/max_terminated_length": 698.0, "completions/mean_length": 450.1875, "completions/mean_terminated_length": 450.1875, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.6268, "grad_norm": 0.08761420845985413, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36863402.0, "reward": 0.7296761274337769, "reward_std": 0.3847619891166687, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27184998989105225, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 462.75, "completions/mean_terminated_length": 462.75, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.627, "grad_norm": 0.14939357340335846, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36875190.0, "reward": 1.2017072439193726, "reward_std": 0.45466142892837524, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7159125208854675, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 433.8125, "completions/mean_terminated_length": 394.4666748046875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.6272, "grad_norm": 0.20379038155078888, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36886115.0, "reward": 1.1975539922714233, "reward_std": 0.49568867683410645, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7701062560081482, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8443750143051147, "step": 3136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 590.9375, "completions/mean_terminated_length": 562.0667114257812, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.6274, "grad_norm": 0.09719612449407578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36899346.0, "reward": 1.2807416915893555, "reward_std": 0.48172008991241455, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8114812970161438, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9131250381469727, "step": 3137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 827.125, "completions/mean_terminated_length": 761.5, "completions/min_length": 514.0, "completions/min_terminated_length": 514.0, "epoch": 0.6276, "grad_norm": 0.10088422894477844, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36916980.0, "reward": 0.5087505578994751, "reward_std": 0.33482229709625244, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26749998331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5625, "step": 3138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 382.5, "completions/mean_terminated_length": 382.5, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.6278, "grad_norm": 0.13536754250526428, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36927644.0, "reward": 0.7454571723937988, "reward_std": 0.3825949430465698, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3034124970436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 3139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 655.5625, "completions/mean_terminated_length": 631.0000610351562, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.628, "grad_norm": 0.04146607965230942, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36941637.0, "reward": 0.5348446369171143, "reward_std": 0.16420964896678925, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13218750059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 3140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 403.1875, "completions/mean_terminated_length": 361.8000183105469, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.6282, "grad_norm": 0.10726647824048996, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36952744.0, "reward": 1.3441102504730225, "reward_std": 0.4294988811016083, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8757187128067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8868750333786011, "step": 3141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 483.0, "completions/max_terminated_length": 483.0, "completions/mean_length": 374.3125, "completions/mean_terminated_length": 374.3125, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.6284, "grad_norm": 0.006214577239006758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36962845.0, "reward": 0.5378947257995605, "reward_std": 0.014474987983703613, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0757874995470047, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 3142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 310.3125, "completions/mean_terminated_length": 310.3125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.6286, "grad_norm": 0.11594240367412567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36971954.0, "reward": 1.4337353706359863, "reward_std": 0.24923662841320038, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9299687743186951, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 3143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 568.0625, "completions/mean_terminated_length": 568.0625, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.6288, "grad_norm": 0.06994480639696121, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36985043.0, "reward": 0.8187506198883057, "reward_std": 0.21084004640579224, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.637499988079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4750000238418579, "step": 3144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 334.9375, "completions/mean_terminated_length": 334.9375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.629, "grad_norm": 0.06552039831876755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36994306.0, "reward": 0.6797663569450378, "reward_std": 0.16313773393630981, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3595312237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7593749761581421, "step": 3145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 392.0, "completions/mean_terminated_length": 392.0, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.6292, "grad_norm": 0.029375022277235985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37004594.0, "reward": 1.4605731964111328, "reward_std": 0.1561535894870758, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9836437106132507, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250619888306, "step": 3146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 410.375, "completions/mean_terminated_length": 410.375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.6294, "grad_norm": 0.23242995142936707, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37017688.0, "reward": 1.0547635555267334, "reward_std": 0.538537323474884, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6095250248908997, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9024999737739563, "step": 3147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 698.4375, "completions/mean_terminated_length": 676.7333984375, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.6296, "grad_norm": 0.05633486807346344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37032927.0, "reward": 0.6767976880073547, "reward_std": 0.2745835781097412, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.41609376668930054, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.590624988079071, "step": 3148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 390.0625, "completions/mean_terminated_length": 390.0625, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.6298, "grad_norm": 0.16598719358444214, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37042848.0, "reward": 0.8053008317947388, "reward_std": 0.37531885504722595, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.42309999465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8237499594688416, "step": 3149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 940.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 699.0, "completions/mean_terminated_length": 699.0, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.63, "grad_norm": 0.0987866148352623, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37059312.0, "reward": 0.9603761434555054, "reward_std": 0.4909863770008087, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4832499921321869, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 3150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 588.0625, "completions/mean_terminated_length": 559.0, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.6302, "grad_norm": 0.11211146414279938, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37073089.0, "reward": 0.8371071219444275, "reward_std": 0.4524402320384979, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4867124855518341, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7587500214576721, "step": 3151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 703.5625, "completions/mean_terminated_length": 657.7857666015625, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.6304, "grad_norm": 0.09749378263950348, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37088442.0, "reward": 0.7324224710464478, "reward_std": 0.38246047496795654, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.52734375, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4593749940395355, "step": 3152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 428.5, "completions/mean_terminated_length": 428.5, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.6306, "grad_norm": 0.12793311476707458, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37099010.0, "reward": 1.3383979797363281, "reward_std": 0.353804349899292, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8642937541007996, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9393750429153442, "step": 3153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 411.0, "completions/mean_terminated_length": 411.0, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.6308, "grad_norm": 0.16342827677726746, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37109714.0, "reward": 1.0455353260040283, "reward_std": 0.4706658720970154, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5910687446594238, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9268749952316284, "step": 3154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 713.9375, "completions/mean_terminated_length": 693.2667236328125, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.631, "grad_norm": 0.057635631412267685, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37126593.0, "reward": 0.884375274181366, "reward_std": 0.24517257511615753, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.831250011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.26249998807907104, "step": 3155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 420.125, "completions/mean_terminated_length": 420.125, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.6312, "grad_norm": 0.09480652213096619, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37137315.0, "reward": 0.8329693675041199, "reward_std": 0.22077251970767975, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6659374833106995, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4312499761581421, "step": 3156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 379.125, "completions/mean_terminated_length": 379.125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.6314, "grad_norm": 0.14553768932819366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37147461.0, "reward": 0.9818729162216187, "reward_std": 0.4781534671783447, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5262437462806702, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 3157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 636.6875, "completions/mean_terminated_length": 610.86669921875, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.6316, "grad_norm": 0.10096222162246704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37162336.0, "reward": 0.8153128623962402, "reward_std": 0.29136499762535095, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6931250095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3374999761581421, "step": 3158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 550.3125, "completions/mean_terminated_length": 550.3125, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.6318, "grad_norm": 0.10838174819946289, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37175397.0, "reward": 1.0837479829788208, "reward_std": 0.486576646566391, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6049937605857849, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 392.5, "completions/mean_terminated_length": 392.5, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.632, "grad_norm": 0.04863704368472099, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37186349.0, "reward": 0.5647542476654053, "reward_std": 0.11250057071447372, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12950624525547028, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143749475479126, "step": 3160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 633.8125, "completions/mean_terminated_length": 633.8125, "completions/min_length": 520.0, "completions/min_terminated_length": 520.0, "epoch": 0.6322, "grad_norm": 0.11205952614545822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37200682.0, "reward": 0.9813636541366577, "reward_std": 0.48491528630256653, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5252249836921692, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262499809265137, "step": 3161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 666.8125, "completions/mean_terminated_length": 643.0000610351562, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.6324, "grad_norm": 0.05066152289509773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37215335.0, "reward": 0.564141571521759, "reward_std": 0.21608133614063263, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19078123569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 3162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 530.0625, "completions/mean_terminated_length": 530.0625, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.6326, "grad_norm": 0.07629554718732834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37227528.0, "reward": 1.3830227851867676, "reward_std": 0.3167521059513092, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8910437822341919, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 3163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 401.5625, "completions/mean_terminated_length": 401.5625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.6328, "grad_norm": 0.14360348880290985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37237777.0, "reward": 0.7807228565216064, "reward_std": 0.42826786637306213, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3114437460899353, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249380111694, "step": 3164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 821.9375, "completions/mean_terminated_length": 664.7777709960938, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.633, "grad_norm": 0.12651795148849487, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37254688.0, "reward": 0.7719535827636719, "reward_std": 0.7323015928268433, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.4814062714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4781250059604645, "step": 3165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 600.3125, "completions/mean_terminated_length": 572.0667114257812, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.6332, "grad_norm": 0.09908010810613632, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37267989.0, "reward": 0.7731256484985352, "reward_std": 0.36263221502304077, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.48374998569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.699999988079071, "step": 3166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 404.25, "completions/mean_terminated_length": 404.25, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.6334, "grad_norm": 0.08932920545339584, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37278425.0, "reward": 0.6080353856086731, "reward_std": 0.23843620717525482, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15356874465942383, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 495.5, "completions/mean_terminated_length": 495.5, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.6336, "grad_norm": 0.12360740453004837, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37290369.0, "reward": 0.8412508964538574, "reward_std": 0.40318262577056885, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4325000047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 3168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 399.0, "completions/mean_terminated_length": 399.0, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.6338, "grad_norm": 0.008184768259525299, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37300321.0, "reward": 0.5292887687683105, "reward_std": 0.023038646206259727, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05857499688863754, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 3169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 349.0625, "completions/mean_terminated_length": 349.0625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.634, "grad_norm": 0.13249829411506653, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37310498.0, "reward": 0.5936699509620667, "reward_std": 0.23125100135803223, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1248374953866005, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 3170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 705.5, "completions/mean_terminated_length": 705.5, "completions/min_length": 548.0, "completions/min_terminated_length": 548.0, "epoch": 0.6342, "grad_norm": 0.004588608630001545, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37326042.0, "reward": 0.5342322587966919, "reward_std": 0.017700711265206337, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06846249848604202, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 3171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 672.5625, "completions/mean_terminated_length": 622.357177734375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.6344, "grad_norm": 0.08167881518602371, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37340739.0, "reward": 0.620703935623169, "reward_std": 0.2876828908920288, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.36640626192092896, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 3172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 863.875, "completions/mean_terminated_length": 791.0909423828125, "completions/min_length": 588.0, "completions/min_terminated_length": 588.0, "epoch": 0.6346, "grad_norm": 0.1662474125623703, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37359361.0, "reward": 0.529531717300415, "reward_std": 0.4328363537788391, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.30906248092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.53125, "step": 3173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 415.125, "completions/mean_terminated_length": 415.125, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.6348, "grad_norm": 0.13771429657936096, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37369843.0, "reward": 0.765313446521759, "reward_std": 0.38094979524612427, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.34312498569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 3174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 431.0, "completions/max_terminated_length": 431.0, "completions/mean_length": 375.875, "completions/mean_terminated_length": 375.875, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.635, "grad_norm": 0.12964577972888947, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37380353.0, "reward": 0.9645165801048279, "reward_std": 0.48689496517181396, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49153125286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 432.9375, "completions/mean_terminated_length": 432.9375, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.6352, "grad_norm": 0.00593282887712121, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37391344.0, "reward": 0.551641583442688, "reward_std": 0.011562498286366463, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10328124463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 795.0625, "completions/mean_terminated_length": 742.2307739257812, "completions/min_length": 534.0, "completions/min_terminated_length": 534.0, "epoch": 0.6354, "grad_norm": 0.10287795215845108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37408065.0, "reward": 0.6017192602157593, "reward_std": 0.3529152274131775, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3909375071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5125000476837158, "step": 3177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 724.875, "completions/mean_terminated_length": 682.1428833007812, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.6356, "grad_norm": 0.08051507920026779, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37424015.0, "reward": 1.047253966331482, "reward_std": 0.5406225323677063, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7195062041282654, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6581250429153442, "step": 3178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 602.375, "completions/mean_terminated_length": 574.2667236328125, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.6358, "grad_norm": 0.04732385277748108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37437317.0, "reward": 0.6910162568092346, "reward_std": 0.23293337225914001, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.44453126192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6468750238418579, "step": 3179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 254.0, "completions/max_terminated_length": 254.0, "completions/mean_length": 192.0, "completions/mean_terminated_length": 192.0, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.636, "grad_norm": 0.34337443113327026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37444357.0, "reward": 0.9002915620803833, "reward_std": 0.4792352020740509, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4255812466144562, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 385.0625, "completions/mean_terminated_length": 385.0625, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.6362, "grad_norm": 0.2069583386182785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37454198.0, "reward": 1.2002822160720825, "reward_std": 0.45774880051612854, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7130625247955322, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 3181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 294.625, "completions/mean_terminated_length": 294.625, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.6364, "grad_norm": 0.23829466104507446, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37462448.0, "reward": 1.140141487121582, "reward_std": 0.47837331891059875, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6552813053131104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 361.4375, "completions/mean_terminated_length": 361.4375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.6366, "grad_norm": 0.24979305267333984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37472727.0, "reward": 1.2271039485931396, "reward_std": 0.5015366077423096, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7667062282562256, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 3183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 689.3125, "completions/mean_terminated_length": 667.0000610351562, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.6368, "grad_norm": 0.044583797454833984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37488556.0, "reward": 0.9023439288139343, "reward_std": 0.2649169862270355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8671875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.16249999403953552, "step": 3184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 439.5, "completions/mean_terminated_length": 439.5, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.637, "grad_norm": 0.1293697953224182, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37499396.0, "reward": 0.9166384339332581, "reward_std": 0.4662172794342041, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4582749903202057, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "step": 3185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 513.875, "completions/mean_terminated_length": 513.875, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.6372, "grad_norm": 0.11837396770715714, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37511650.0, "reward": 0.6507695913314819, "reward_std": 0.24932429194450378, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23903748393058777, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 357.4375, "completions/mean_terminated_length": 357.4375, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.6374, "grad_norm": 0.20407918095588684, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37520793.0, "reward": 0.9567697644233704, "reward_std": 0.4945070445537567, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47603750228881836, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 3187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 320.25, "completions/mean_terminated_length": 320.25, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.6376, "grad_norm": 0.012383605353534222, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37529901.0, "reward": 0.5200420022010803, "reward_std": 0.026967236772179604, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04008124768733978, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 3188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 435.25, "completions/mean_terminated_length": 435.25, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.6378, "grad_norm": 0.11909998208284378, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37541105.0, "reward": 0.784507155418396, "reward_std": 0.42610934376716614, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31901249289512634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 637.75, "completions/mean_terminated_length": 637.75, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.638, "grad_norm": 0.10967359691858292, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37555357.0, "reward": 0.8409197330474854, "reward_std": 0.4585545063018799, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3693374693393707, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 3190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 639.4375, "completions/mean_terminated_length": 639.4375, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.6382, "grad_norm": 0.11376404762268066, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37569492.0, "reward": 0.9066009521484375, "reward_std": 0.4762735366821289, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4381999969482422, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 600.625, "completions/mean_terminated_length": 540.1428833007812, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.6384, "grad_norm": 0.1060524433851242, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37583294.0, "reward": 0.6707819700241089, "reward_std": 0.4176467955112457, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.34156250953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.699999988079071, "step": 3192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 491.875, "completions/mean_terminated_length": 456.4000244140625, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.6386, "grad_norm": 0.07200832664966583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37595548.0, "reward": 0.4728759527206421, "reward_std": 0.18526771664619446, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07074999809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 3193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 296.0, "completions/mean_terminated_length": 296.0, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.6388, "grad_norm": 0.06473097205162048, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37604524.0, "reward": 1.4395854473114014, "reward_std": 0.23755626380443573, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416687488555908, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 3194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 477.875, "completions/mean_terminated_length": 477.875, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.639, "grad_norm": 0.10217422246932983, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37616314.0, "reward": 0.6728917360305786, "reward_std": 0.32433873414993286, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22078125178813934, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 710.0, "completions/max_terminated_length": 710.0, "completions/mean_length": 484.625, "completions/mean_terminated_length": 484.625, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.6392, "grad_norm": 0.11811739951372147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37627876.0, "reward": 1.1032729148864746, "reward_std": 0.4188968241214752, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7065437436103821, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7793750166893005, "step": 3196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 658.25, "completions/mean_terminated_length": 658.25, "completions/min_length": 485.0, "completions/min_terminated_length": 485.0, "epoch": 0.6394, "grad_norm": 0.07899439334869385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37643160.0, "reward": 0.7603914737701416, "reward_std": 0.3828028738498688, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3332812488079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 3197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 744.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 534.5, "completions/mean_terminated_length": 534.5, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.6396, "grad_norm": 0.1257280707359314, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37655696.0, "reward": 0.8317975997924805, "reward_std": 0.298848420381546, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5385937690734863, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7093749642372131, "step": 3198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 424.75, "completions/mean_terminated_length": 424.75, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.6398, "grad_norm": 0.1343957781791687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37666284.0, "reward": 0.9311727285385132, "reward_std": 0.4553770124912262, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4873437285423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 486.6875, "completions/mean_terminated_length": 486.6875, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.64, "grad_norm": 0.1330496221780777, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37678663.0, "reward": 1.1892449855804443, "reward_std": 0.4758065640926361, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.690987229347229, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9898749589920044, "step": 3200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 563.0, "completions/mean_terminated_length": 532.2667236328125, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.6402, "grad_norm": 0.16525577008724213, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37691655.0, "reward": 0.7798445820808411, "reward_std": 0.4658896327018738, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3721874952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 3201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 592.5625, "completions/mean_terminated_length": 530.9285888671875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.6404, "grad_norm": 0.045045141130685806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37706272.0, "reward": 0.5147663950920105, "reward_std": 0.22120583057403564, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15453124046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.784375011920929, "step": 3202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 616.125, "completions/mean_terminated_length": 616.125, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.6406, "grad_norm": 0.04437165707349777, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37720274.0, "reward": 0.8192974925041199, "reward_std": 0.2261561155319214, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5760937929153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6218750476837158, "step": 3203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 612.0, "completions/mean_terminated_length": 612.0, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.6408, "grad_norm": 0.11369562149047852, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37733986.0, "reward": 1.261535406112671, "reward_std": 0.42502281069755554, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7730687856674194, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 652.375, "completions/mean_terminated_length": 652.375, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.641, "grad_norm": 0.027532700449228287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37748648.0, "reward": 0.8783597946166992, "reward_std": 0.12291975319385529, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7567187547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.41562503576278687, "step": 3205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 461.625, "completions/mean_terminated_length": 461.625, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.6412, "grad_norm": 0.10852108150720596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37760098.0, "reward": 0.7782729864120483, "reward_std": 0.4304492473602295, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30654376745224, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 3206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 266.0, "completions/max_terminated_length": 266.0, "completions/mean_length": 222.5625, "completions/mean_terminated_length": 222.5625, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.6414, "grad_norm": 0.11936726421117783, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37767579.0, "reward": 0.6081916093826294, "reward_std": 0.23776249587535858, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1538812518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 545.875, "completions/mean_terminated_length": 545.875, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.6416, "grad_norm": 0.12472528964281082, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37780137.0, "reward": 0.8021729588508606, "reward_std": 0.42181363701820374, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3543437719345093, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9293749928474426, "step": 3208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 299.9375, "completions/mean_terminated_length": 299.9375, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.6418, "grad_norm": 0.20568054914474487, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37788376.0, "reward": 1.3206260204315186, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 408.625, "completions/mean_terminated_length": 408.625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.642, "grad_norm": 0.3719448447227478, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37798994.0, "reward": 1.0782040357589722, "reward_std": 0.4826279878616333, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5939062833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 3210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 391.1875, "completions/mean_terminated_length": 391.1875, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.6422, "grad_norm": 0.2257164865732193, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37809141.0, "reward": 0.9626321792602539, "reward_std": 0.4885059893131256, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4877625107765198, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 450.9375, "completions/mean_terminated_length": 412.7333679199219, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.6424, "grad_norm": 0.15862661600112915, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37819956.0, "reward": 0.7753133773803711, "reward_std": 0.422284334897995, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4256249964237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.762499988079071, "step": 3212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 720.75, "completions/mean_terminated_length": 650.7692260742188, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.6426, "grad_norm": 0.16731512546539307, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37835344.0, "reward": 0.7426571249961853, "reward_std": 0.5669322609901428, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36031249165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 3213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 713.0, "completions/mean_terminated_length": 692.2667236328125, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.6428, "grad_norm": 0.096552774310112, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37852624.0, "reward": 0.7743631601333618, "reward_std": 0.3768140375614166, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4862250089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6425000429153442, "step": 3214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 636.9375, "completions/mean_terminated_length": 581.6428833007812, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.643, "grad_norm": 0.1866331696510315, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37866671.0, "reward": 1.012528896331787, "reward_std": 0.5942868590354919, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5875562429428101, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.846875011920929, "step": 3215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 321.5, "completions/mean_terminated_length": 321.5, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.6432, "grad_norm": 0.18345943093299866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37875511.0, "reward": 0.9581229090690613, "reward_std": 0.4923451244831085, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4787437617778778, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 993.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 487.0, "completions/mean_terminated_length": 487.0, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.6434, "grad_norm": 0.08646609634160995, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37887175.0, "reward": 0.8055475950241089, "reward_std": 0.20166492462158203, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6110937595367432, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 3217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 329.75, "completions/mean_terminated_length": 329.75, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.6436, "grad_norm": 0.2196800708770752, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37896819.0, "reward": 1.3800759315490723, "reward_std": 0.3245161771774292, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851500153541565, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 830.0, "completions/mean_terminated_length": 741.8181762695312, "completions/min_length": 543.0, "completions/min_terminated_length": 543.0, "epoch": 0.6438, "grad_norm": 0.11700109392404556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37913859.0, "reward": 0.5231255292892456, "reward_std": 0.4237983524799347, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23375000059604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.39374998211860657, "step": 3219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 371.25, "completions/mean_terminated_length": 371.25, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.644, "grad_norm": 0.0001927033154061064, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37923751.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 622.875, "completions/mean_terminated_length": 622.875, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.6442, "grad_norm": 0.09527380019426346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37937573.0, "reward": 0.699050784111023, "reward_std": 0.26370441913604736, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.33560001850128174, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8112499713897705, "step": 3221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 422.625, "completions/mean_terminated_length": 422.625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.6444, "grad_norm": 0.2988389730453491, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37947935.0, "reward": 1.0831633806228638, "reward_std": 0.4867655634880066, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6038249731063843, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 3222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 299.0, "completions/max_terminated_length": 299.0, "completions/mean_length": 258.9375, "completions/mean_terminated_length": 258.9375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.6446, "grad_norm": 0.0002509189653210342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37956158.0, "reward": 1.4996823072433472, "reward_std": 0.000467625999590382, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9993625283241272, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 3223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 710.0, "completions/max_terminated_length": 710.0, "completions/mean_length": 425.8125, "completions/mean_terminated_length": 425.8125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.6448, "grad_norm": 0.1300022453069687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37966683.0, "reward": 1.0388164520263672, "reward_std": 0.4658184349536896, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5776312351226807, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 3224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 704.125, "completions/mean_terminated_length": 682.800048828125, "completions/min_length": 450.0, "completions/min_terminated_length": 450.0, "epoch": 0.645, "grad_norm": 0.0847146213054657, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37983469.0, "reward": 0.6723445057868958, "reward_std": 0.2689850926399231, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.40718749165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.625, "step": 3225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 744.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 516.5625, "completions/mean_terminated_length": 516.5625, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.6452, "grad_norm": 0.07693034410476685, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37995638.0, "reward": 1.3539073467254639, "reward_std": 0.3276471495628357, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.832812488079071, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 3226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 346.625, "completions/mean_terminated_length": 346.625, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.6454, "grad_norm": 0.15126270055770874, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38004880.0, "reward": 1.3207571506500244, "reward_std": 0.3830225169658661, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290125131607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 395.5625, "completions/mean_terminated_length": 395.5625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.6456, "grad_norm": 0.09608469158411026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38015033.0, "reward": 0.6602197885513306, "reward_std": 0.32771074771881104, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.195437490940094, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 367.0625, "completions/mean_terminated_length": 367.0625, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.6458, "grad_norm": 0.2078295797109604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38024906.0, "reward": 0.842579185962677, "reward_std": 0.45705699920654297, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3726562559604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 529.375, "completions/mean_terminated_length": 496.4000244140625, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.646, "grad_norm": 0.16200889647006989, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38037616.0, "reward": 0.9801384210586548, "reward_std": 0.5373746752738953, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5227749943733215, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8725000619888306, "step": 3230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 400.0, "completions/mean_terminated_length": 400.0, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6462, "grad_norm": 0.21551717817783356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38048256.0, "reward": 0.8971697092056274, "reward_std": 0.4816673994064331, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41933751106262207, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 3231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 698.5, "completions/mean_terminated_length": 623.3846435546875, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.6464, "grad_norm": 0.1252160668373108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38064056.0, "reward": 0.8950788378715515, "reward_std": 0.6046519875526428, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5401562452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 638.0, "completions/mean_terminated_length": 509.3333435058594, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.6466, "grad_norm": 0.08941831439733505, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38078424.0, "reward": 0.40247267484664917, "reward_std": 0.240514874458313, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05494374781847, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.721875011920929, "step": 3233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 605.8125, "completions/mean_terminated_length": 605.8125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.6468, "grad_norm": 0.10715948045253754, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38091973.0, "reward": 0.7579693794250488, "reward_std": 0.2655424475669861, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5784375071525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.48124998807907104, "step": 3234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 491.0, "completions/max_terminated_length": 491.0, "completions/mean_length": 415.875, "completions/mean_terminated_length": 415.875, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.647, "grad_norm": 0.2544647455215454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38102403.0, "reward": 1.2047665119171143, "reward_std": 0.45047640800476074, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7220312356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 956.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 594.125, "completions/mean_terminated_length": 594.125, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.6472, "grad_norm": 0.1359577476978302, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38115589.0, "reward": 0.6996451616287231, "reward_std": 0.3972647786140442, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21178750693798065, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 3236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 370.375, "completions/mean_terminated_length": 370.375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.6474, "grad_norm": 0.25585922598838806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38124923.0, "reward": 1.0667572021484375, "reward_std": 0.491829514503479, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5710124969482422, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 3237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 764.5, "completions/mean_terminated_length": 747.2000122070312, "completions/min_length": 529.0, "completions/min_terminated_length": 529.0, "epoch": 0.6476, "grad_norm": 0.11332349479198456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38141171.0, "reward": 0.8528663516044617, "reward_std": 0.4433586299419403, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.518231213092804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7443749904632568, "step": 3238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 313.8125, "completions/mean_terminated_length": 313.8125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.6478, "grad_norm": 0.2584754228591919, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38149904.0, "reward": 1.375897765159607, "reward_std": 0.3361135423183441, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8767937421798706, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 329.5625, "completions/mean_terminated_length": 329.5625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.648, "grad_norm": 0.1129288598895073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38159113.0, "reward": 1.3161543607711792, "reward_std": 0.39369067549705505, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8198062777519226, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 3240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 506.75, "completions/mean_terminated_length": 506.75, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.6482, "grad_norm": 0.22592833638191223, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38171093.0, "reward": 1.2517977952957153, "reward_std": 0.4100543260574341, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7535937428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 3241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 918.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 592.3125, "completions/mean_terminated_length": 592.3125, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.6484, "grad_norm": 0.134761244058609, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38184154.0, "reward": 1.2167850732803345, "reward_std": 0.3789359927177429, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7460687160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7706249952316284, "step": 3242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 975.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 703.5, "completions/mean_terminated_length": 703.5, "completions/min_length": 435.0, "completions/min_terminated_length": 435.0, "epoch": 0.6486, "grad_norm": 0.13361017405986786, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38199970.0, "reward": 0.9021633267402649, "reward_std": 0.43347159028053284, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4918249845504761, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8725000619888306, "step": 3243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 710.6875, "completions/mean_terminated_length": 638.3846435546875, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.6488, "grad_norm": 0.04967927560210228, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38215901.0, "reward": 0.4960947036743164, "reward_std": 0.3433802127838135, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1171875, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.78125, "step": 3244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 244.125, "completions/mean_terminated_length": 244.125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.649, "grad_norm": 0.33351826667785645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38223359.0, "reward": 0.7791103720664978, "reward_std": 0.42972472310066223, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.308218777179718, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 3245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 488.1875, "completions/mean_terminated_length": 488.1875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.6492, "grad_norm": 0.11896196007728577, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38234754.0, "reward": 0.6018761396408081, "reward_std": 0.24036246538162231, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14124999940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 500.375, "completions/mean_terminated_length": 500.375, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.6494, "grad_norm": 0.03933101147413254, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38247368.0, "reward": 1.4400166273117065, "reward_std": 0.23767150938510895, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9425312876701355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 3247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 500.0, "completions/mean_terminated_length": 500.0, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.6496, "grad_norm": 0.13217014074325562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38259144.0, "reward": 1.3210384845733643, "reward_std": 0.3831622302532196, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8295750021934509, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 3248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 362.1875, "completions/mean_terminated_length": 362.1875, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.6498, "grad_norm": 0.21634918451309204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38268427.0, "reward": 0.8468383550643921, "reward_std": 0.4544425308704376, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3811750113964081, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 330.5, "completions/mean_terminated_length": 330.5, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.65, "grad_norm": 0.0028137017507106066, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38277843.0, "reward": 0.5469415187835693, "reward_std": 0.007237497717142105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09388124942779541, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 852.0, "completions/mean_terminated_length": 812.3077392578125, "completions/min_length": 626.0, "completions/min_terminated_length": 626.0, "epoch": 0.6502, "grad_norm": 0.208090141415596, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38295315.0, "reward": 0.6666256189346313, "reward_std": 0.4617297351360321, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3957499861717224, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5824999809265137, "step": 3251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 239.0, "completions/max_terminated_length": 239.0, "completions/mean_length": 194.4375, "completions/mean_terminated_length": 194.4375, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.6504, "grad_norm": 0.6179668307304382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38302234.0, "reward": 0.8336166143417358, "reward_std": 0.46352750062942505, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3547312617301941, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 3252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 583.6875, "completions/mean_terminated_length": 583.6875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.6506, "grad_norm": 0.07213421911001205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38315829.0, "reward": 0.5950759649276733, "reward_std": 0.15428034961223602, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19014999270439148, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 3253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 720.0, "completions/mean_terminated_length": 676.5714721679688, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.6508, "grad_norm": 0.11535827815532684, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38331221.0, "reward": 0.7735788226127625, "reward_std": 0.46973860263824463, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4846562445163727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6106249690055847, "step": 3254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 821.375, "completions/mean_terminated_length": 699.7999877929688, "completions/min_length": 464.0, "completions/min_terminated_length": 464.0, "epoch": 0.651, "grad_norm": 0.06979841738939285, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38348299.0, "reward": 0.3837505280971527, "reward_std": 0.3279065489768982, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14249998331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5249999761581421, "step": 3255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 799.875, "completions/mean_terminated_length": 665.4000244140625, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.6512, "grad_norm": 0.15249086916446686, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38365737.0, "reward": 0.42834433913230896, "reward_std": 0.4952491819858551, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1691874861717224, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5325000286102295, "step": 3256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 320.6875, "completions/mean_terminated_length": 320.6875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.6514, "grad_norm": 0.22584031522274017, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38374468.0, "reward": 0.7736511826515198, "reward_std": 0.43312665820121765, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2973000109195709, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 3257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 465.3125, "completions/mean_terminated_length": 465.3125, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.6516, "grad_norm": 0.15634338557720184, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38386633.0, "reward": 0.9445977210998535, "reward_std": 0.45744946599006653, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5141937732696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8831249475479126, "step": 3258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 273.0, "completions/max_terminated_length": 273.0, "completions/mean_length": 215.3125, "completions/mean_terminated_length": 215.3125, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.6518, "grad_norm": 0.10220382362604141, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38394766.0, "reward": 0.5947573184967041, "reward_std": 0.24192336201667786, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1270124912261963, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 3259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 797.0625, "completions/mean_terminated_length": 781.933349609375, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.652, "grad_norm": 0.15105053782463074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38411807.0, "reward": 0.6238292455673218, "reward_std": 0.36660221219062805, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18515625596046448, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 3260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 369.5, "completions/mean_terminated_length": 369.5, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.6522, "grad_norm": 0.2193567454814911, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38421367.0, "reward": 1.1906228065490723, "reward_std": 0.4254120886325836, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7562437653541565, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8681250214576721, "step": 3261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 473.375, "completions/mean_terminated_length": 436.66668701171875, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.6524, "grad_norm": 0.2553862929344177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38433453.0, "reward": 1.0099761486053467, "reward_std": 0.536322832107544, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5824500322341919, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8600000143051147, "step": 3262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 345.375, "completions/mean_terminated_length": 345.375, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.6526, "grad_norm": 0.03393884375691414, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38443155.0, "reward": 0.5649977922439575, "reward_std": 0.054221261292696, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.129993736743927, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306250214576721, "step": 3263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 292.375, "completions/mean_terminated_length": 292.375, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.6528, "grad_norm": 0.0010295098181813955, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38451961.0, "reward": 0.9987507462501526, "reward_std": 0.00223643914796412, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.02500000037252903, "step": 3264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 291.875, "completions/mean_terminated_length": 291.875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.653, "grad_norm": 0.1987454891204834, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38460279.0, "reward": 0.8877226710319519, "reward_std": 0.32168006896972656, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6504437327384949, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5206249952316284, "step": 3265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 512.5, "completions/mean_terminated_length": 512.5, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.6532, "grad_norm": 0.07189016789197922, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38473167.0, "reward": 0.6298414468765259, "reward_std": 0.16245010495185852, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25968125462532043, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8118749856948853, "step": 3266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 434.5625, "completions/mean_terminated_length": 434.5625, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.6534, "grad_norm": 0.2378440499305725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38483848.0, "reward": 0.7297666072845459, "reward_std": 0.38169553875923157, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2720312476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 663.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 480.8125, "completions/mean_terminated_length": 480.8125, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.6536, "grad_norm": 0.1393355131149292, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38496309.0, "reward": 0.8019539713859558, "reward_std": 0.37510713934898376, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4164062738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8343750238418579, "step": 3268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 307.625, "completions/mean_terminated_length": 307.625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.6538, "grad_norm": 0.23637908697128296, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38505295.0, "reward": 1.2023135423660278, "reward_std": 0.45508289337158203, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7171249985694885, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 3269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 356.8125, "completions/mean_terminated_length": 356.8125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.654, "grad_norm": 0.21365085244178772, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38514668.0, "reward": 0.899372935295105, "reward_std": 0.4793868660926819, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42374375462532043, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 349.5625, "completions/mean_terminated_length": 349.5625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.6542, "grad_norm": 0.0003344444849062711, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38523957.0, "reward": 1.4992542266845703, "reward_std": 0.0005924947909079492, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985062479972839, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 3271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 615.75, "completions/mean_terminated_length": 615.75, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.6544, "grad_norm": 0.08520112186670303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38537873.0, "reward": 0.7683475017547607, "reward_std": 0.20842449367046356, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5366937518119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5893750190734863, "step": 3272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 761.625, "completions/mean_terminated_length": 744.1333618164062, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.6546, "grad_norm": 0.09651445597410202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38553931.0, "reward": 0.5924224853515625, "reward_std": 0.28946158289909363, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3098437190055847, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6531249284744263, "step": 3273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 308.5, "completions/mean_terminated_length": 260.8000183105469, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.6548, "grad_norm": 0.3715774416923523, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38562563.0, "reward": 0.7471789717674255, "reward_std": 0.4680769443511963, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.306856244802475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981250524520874, "step": 3274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 494.625, "completions/mean_terminated_length": 494.625, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.655, "grad_norm": 0.10883629322052002, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38574541.0, "reward": 0.6361727118492126, "reward_std": 0.17771302163600922, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20984375476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 3275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 533.5, "completions/mean_terminated_length": 533.5, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.6552, "grad_norm": 0.15609526634216309, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38588085.0, "reward": 1.1088887453079224, "reward_std": 0.46961525082588196, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6552749872207642, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287500381469727, "step": 3276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 760.0625, "completions/mean_terminated_length": 640.0909423828125, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.6554, "grad_norm": 0.3010751008987427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38604614.0, "reward": 0.6058596968650818, "reward_std": 0.4339473247528076, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5242187976837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.3031250238418579, "step": 3277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 248.0, "completions/max_terminated_length": 248.0, "completions/mean_length": 165.625, "completions/mean_terminated_length": 165.625, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "epoch": 0.6556, "grad_norm": 0.17894189059734344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38610848.0, "reward": 1.4399728775024414, "reward_std": 0.2376599758863449, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424437284469604, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706249833106995, "step": 3278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 267.8125, "completions/mean_terminated_length": 267.8125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.6558, "grad_norm": 0.00029564060969278216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38618989.0, "reward": 1.499454140663147, "reward_std": 0.0005658540176227689, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989062547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9718750715255737, "step": 3279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 475.75, "completions/mean_terminated_length": 475.75, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.656, "grad_norm": 0.20804937183856964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38630745.0, "reward": 0.8050007820129395, "reward_std": 0.3786439001560211, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.42249998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187500238418579, "step": 3280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 321.6875, "completions/mean_terminated_length": 321.6875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.6562, "grad_norm": 0.17570187151432037, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38639524.0, "reward": 1.3770197629928589, "reward_std": 0.3329569399356842, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8790374994277954, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 3281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 413.0, "completions/mean_terminated_length": 413.0, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.6564, "grad_norm": 0.10631964355707169, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38649924.0, "reward": 0.6577947735786438, "reward_std": 0.32873255014419556, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19058750569820404, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 3282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 456.0, "completions/mean_terminated_length": 418.13336181640625, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.6566, "grad_norm": 0.29711419343948364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38661172.0, "reward": 0.8489853739738464, "reward_std": 0.5553721189498901, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4479687213897705, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 3283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 494.5, "completions/mean_terminated_length": 494.5, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.6568, "grad_norm": 0.23199664056301117, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38673068.0, "reward": 1.0337352752685547, "reward_std": 0.483448326587677, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5674687623977661, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 3284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 300.125, "completions/mean_terminated_length": 251.86668395996094, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.657, "grad_norm": 0.1783822774887085, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38681870.0, "reward": 1.3116321563720703, "reward_std": 0.40309759974479675, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.9357625246047974, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7325000166893005, "step": 3285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 347.5625, "completions/mean_terminated_length": 347.5625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.6572, "grad_norm": 0.03173099085688591, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38691991.0, "reward": 0.5875009298324585, "reward_std": 0.10849516093730927, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17499999701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.893750011920929, "step": 3286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 445.75, "completions/mean_terminated_length": 445.75, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.6574, "grad_norm": 0.16778966784477234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38703459.0, "reward": 0.8267731666564941, "reward_std": 0.46828171610832214, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34104374051094055, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 3287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 900.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 634.0625, "completions/mean_terminated_length": 634.0625, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.6576, "grad_norm": 0.021441103890538216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38717588.0, "reward": 0.5864071846008301, "reward_std": 0.06687804311513901, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17281249165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9062499403953552, "step": 3288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 420.625, "completions/mean_terminated_length": 420.625, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.6578, "grad_norm": 0.20584185421466827, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38728622.0, "reward": 1.20164155960083, "reward_std": 0.45461562275886536, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7157812714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 453.1875, "completions/mean_terminated_length": 453.1875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.658, "grad_norm": 0.08842954784631729, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38740209.0, "reward": 1.3780105113983154, "reward_std": 0.32390087842941284, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8810187578201294, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 3290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 438.0, "completions/max_terminated_length": 438.0, "completions/mean_length": 378.6875, "completions/mean_terminated_length": 378.6875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.6582, "grad_norm": 0.22872664034366608, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38749948.0, "reward": 1.1366041898727417, "reward_std": 0.48350098729133606, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6482062339782715, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 376.6875, "completions/mean_terminated_length": 376.6875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.6584, "grad_norm": 0.22006623446941376, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38759767.0, "reward": 0.7710168957710266, "reward_std": 0.4344688355922699, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2920312285423279, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 3292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 506.4375, "completions/mean_terminated_length": 506.4375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.6586, "grad_norm": 0.15720121562480927, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38772046.0, "reward": 1.234844446182251, "reward_std": 0.32500943541526794, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.9071874618530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6937500238418579, "step": 3293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 294.25, "completions/mean_terminated_length": 294.25, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.6588, "grad_norm": 0.25587198138237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38781394.0, "reward": 0.8420822024345398, "reward_std": 0.45751258730888367, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3716624975204468, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 578.75, "completions/mean_terminated_length": 549.0667114257812, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.659, "grad_norm": 0.11012057214975357, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38794478.0, "reward": 0.4827353358268738, "reward_std": 0.19026324152946472, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09046874940395355, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.828125, "step": 3295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 396.5, "completions/mean_terminated_length": 396.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.6592, "grad_norm": 0.2334088534116745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38804502.0, "reward": 0.966657280921936, "reward_std": 0.4858883321285248, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4958125352859497, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 3296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 313.5625, "completions/mean_terminated_length": 313.5625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.6594, "grad_norm": 0.008523587137460709, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38813423.0, "reward": 0.5455977916717529, "reward_std": 0.01855016127228737, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09119375050067902, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 491.375, "completions/mean_terminated_length": 491.375, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.6596, "grad_norm": 0.14504584670066833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38826293.0, "reward": 1.1455321311950684, "reward_std": 0.47138455510139465, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6660624742507935, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 3298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 361.1875, "completions/mean_terminated_length": 361.1875, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.6598, "grad_norm": 0.30945897102355957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38835624.0, "reward": 1.162172794342041, "reward_std": 0.5300266146659851, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6993438005447388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981250524520874, "step": 3299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 556.8125, "completions/mean_terminated_length": 525.6666870117188, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.66, "grad_norm": 0.0941389948129654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38848949.0, "reward": 1.2865352630615234, "reward_std": 0.4712175726890564, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8230687975883484, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 626.1875, "completions/mean_terminated_length": 626.1875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.6602, "grad_norm": 0.14018817245960236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38863144.0, "reward": 1.046015977859497, "reward_std": 0.29073768854141235, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8420312404632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4281249940395355, "step": 3301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 285.25, "completions/mean_terminated_length": 285.25, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.6604, "grad_norm": 0.09264300763607025, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38871196.0, "reward": 1.4399197101593018, "reward_std": 0.23764562606811523, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9423375129699707, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 3302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 320.8125, "completions/mean_terminated_length": 320.8125, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.6606, "grad_norm": 0.23907285928726196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38880121.0, "reward": 1.0238165855407715, "reward_std": 0.49064570665359497, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5476312637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 395.0625, "completions/mean_terminated_length": 395.0625, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.6608, "grad_norm": 0.212095245718956, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38890698.0, "reward": 1.019441843032837, "reward_std": 0.4962973892688751, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.538881242275238, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756249785423279, "step": 3304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 731.0, "completions/mean_terminated_length": 663.3846435546875, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.661, "grad_norm": 0.1525508314371109, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38906474.0, "reward": 0.8867976665496826, "reward_std": 0.6085587739944458, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5235937833786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.734375, "step": 3305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 800.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 568.6875, "completions/mean_terminated_length": 568.6875, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.6612, "grad_norm": 0.0596553236246109, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38919989.0, "reward": 0.7200008630752563, "reward_std": 0.25153347849845886, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3774999976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.800000011920929, "step": 3306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 255.625, "completions/mean_terminated_length": 255.625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.6614, "grad_norm": 0.26369136571884155, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38927743.0, "reward": 1.21832275390625, "reward_std": 0.43756139278411865, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7491437196731567, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9356250166893005, "step": 3307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 338.125, "completions/mean_terminated_length": 338.125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.6616, "grad_norm": 0.226448655128479, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38936833.0, "reward": 0.8396478295326233, "reward_std": 0.45925605297088623, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36679375171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 392.125, "completions/mean_terminated_length": 392.125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.6618, "grad_norm": 0.16758733987808228, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38947763.0, "reward": 0.9049760103225708, "reward_std": 0.47561654448509216, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43494999408721924, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 274.1875, "completions/mean_terminated_length": 274.1875, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.662, "grad_norm": 0.19050909578800201, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38955622.0, "reward": 0.7124009728431702, "reward_std": 0.39075249433517456, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23730000853538513, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 3310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 390.8125, "completions/mean_terminated_length": 390.8125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.6622, "grad_norm": 0.06134819611907005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38965651.0, "reward": 1.4369981288909912, "reward_std": 0.24986632168293, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9364937543869019, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750667572021, "step": 3311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 480.3125, "completions/mean_terminated_length": 444.0666809082031, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.6624, "grad_norm": 0.056427206844091415, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38977112.0, "reward": 0.6271756887435913, "reward_std": 0.24001434445381165, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3168500065803528, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7174999713897705, "step": 3312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 516.4375, "completions/mean_terminated_length": 516.4375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.6626, "grad_norm": 0.31166866421699524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38989583.0, "reward": 0.9656260013580322, "reward_std": 0.45003050565719604, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5562500357627869, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 3313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 305.9375, "completions/mean_terminated_length": 305.9375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.6628, "grad_norm": 0.1190287321805954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38998622.0, "reward": 1.4398949146270752, "reward_std": 0.2376391440629959, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422875046730042, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 3314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 774.4375, "completions/mean_terminated_length": 716.84619140625, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.663, "grad_norm": 0.08582118153572083, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39015013.0, "reward": 0.6322507262229919, "reward_std": 0.4852304458618164, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2644999921321869, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7637500166893005, "step": 3315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 636.375, "completions/mean_terminated_length": 610.5333862304688, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.6632, "grad_norm": 0.0864495187997818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39029883.0, "reward": 0.6758602261543274, "reward_std": 0.27797263860702515, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4142187237739563, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5843750238418579, "step": 3316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 630.5625, "completions/mean_terminated_length": 574.357177734375, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.6634, "grad_norm": 0.030786586925387383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39043940.0, "reward": 0.4739072918891907, "reward_std": 0.18652944266796112, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07281249761581421, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 3317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 581.5625, "completions/mean_terminated_length": 581.5625, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.6636, "grad_norm": 0.11040804535150528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39057373.0, "reward": 1.0054694414138794, "reward_std": 0.39305105805397034, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.635937511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7249999642372131, "step": 3318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 328.75, "completions/mean_terminated_length": 328.75, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.6638, "grad_norm": 0.2119133174419403, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39066649.0, "reward": 1.2076447010040283, "reward_std": 0.44820812344551086, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7277874946594238, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 3319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 857.6875, "completions/mean_terminated_length": 728.3333129882812, "completions/min_length": 596.0, "completions/min_terminated_length": 596.0, "epoch": 0.664, "grad_norm": 0.15742076933383942, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39084724.0, "reward": 0.6453787088394165, "reward_std": 0.6654723286628723, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4157562255859375, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4893750250339508, "step": 3320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 864.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 537.875, "completions/mean_terminated_length": 537.875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.6642, "grad_norm": 0.07886513322591782, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39097714.0, "reward": 0.7169538736343384, "reward_std": 0.14720983803272247, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.433906227350235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 335.9375, "completions/mean_terminated_length": 335.9375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.6644, "grad_norm": 0.175567165017128, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39106961.0, "reward": 0.8409291505813599, "reward_std": 0.45856890082359314, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.369356244802475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 3322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 677.875, "completions/mean_terminated_length": 654.800048828125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.6646, "grad_norm": 0.05912311002612114, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39122495.0, "reward": 0.6002225875854492, "reward_std": 0.21311505138874054, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26294374465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7643749713897705, "step": 3323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 528.25, "completions/mean_terminated_length": 528.25, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.6648, "grad_norm": 0.11852017045021057, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39134755.0, "reward": 1.285860300064087, "reward_std": 0.3839268088340759, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8217188119888306, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 3324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 804.875, "completions/mean_terminated_length": 731.8333740234375, "completions/min_length": 555.0, "completions/min_terminated_length": 555.0, "epoch": 0.665, "grad_norm": 0.13259924948215485, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39151473.0, "reward": 0.6842848658561707, "reward_std": 0.4881398677825928, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4935687780380249, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.47062501311302185, "step": 3325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 692.0625, "completions/mean_terminated_length": 692.0625, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.6652, "grad_norm": 0.07497861981391907, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39166850.0, "reward": 0.7671881914138794, "reward_std": 0.2003251165151596, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5343749523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5687499642372131, "step": 3326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 697.125, "completions/mean_terminated_length": 697.125, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.6654, "grad_norm": 0.08654220402240753, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39181956.0, "reward": 0.8361726403236389, "reward_std": 0.360363632440567, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.48484373092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.796875, "step": 3327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 638.6875, "completions/mean_terminated_length": 613.0000610351562, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.6656, "grad_norm": 0.07940870523452759, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39195919.0, "reward": 0.5825789570808411, "reward_std": 0.21934911608695984, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2276562601327896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.784375011920929, "step": 3328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 740.1875, "completions/mean_terminated_length": 699.6428833007812, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.6658, "grad_norm": 0.15293332934379578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39211650.0, "reward": 0.8074228763580322, "reward_std": 0.529168426990509, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4273437261581421, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.796875, "step": 3329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 770.1875, "completions/mean_terminated_length": 733.9285888671875, "completions/min_length": 535.0, "completions/min_terminated_length": 535.0, "epoch": 0.666, "grad_norm": 0.11766134202480316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39228325.0, "reward": 0.8859911561012268, "reward_std": 0.4771078824996948, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6469812393188477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5631250143051147, "step": 3330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 355.875, "completions/mean_terminated_length": 355.875, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.6662, "grad_norm": 0.1832522749900818, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39237763.0, "reward": 0.5966854095458984, "reward_std": 0.24153421819210052, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13086873292922974, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 520.4375, "completions/mean_terminated_length": 520.4375, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.6664, "grad_norm": 0.060215502977371216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39250266.0, "reward": 0.9217191934585571, "reward_std": 0.20749785006046295, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.7809374928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4312500059604645, "step": 3332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 457.625, "completions/mean_terminated_length": 419.86669921875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6666, "grad_norm": 0.11509492248296738, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39261300.0, "reward": 0.611641526222229, "reward_std": 0.22231616079807281, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.28578126430511475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7406250238418579, "step": 3333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 530.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 336.1875, "completions/mean_terminated_length": 336.1875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.6668, "grad_norm": 0.01833159662783146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39270407.0, "reward": 0.5341824293136597, "reward_std": 0.020408811047673225, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06836249679327011, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 3334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 471.0, "completions/max_terminated_length": 471.0, "completions/mean_length": 374.8125, "completions/mean_terminated_length": 374.8125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.667, "grad_norm": 0.12184476107358932, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39280436.0, "reward": 0.6026977896690369, "reward_std": 0.23922792077064514, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428937464952469, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 386.8125, "completions/mean_terminated_length": 386.8125, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.6672, "grad_norm": 0.17007195949554443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39290545.0, "reward": 0.7704696655273438, "reward_std": 0.3698163330554962, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.35343748331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 3336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 512.1875, "completions/mean_terminated_length": 512.1875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.6674, "grad_norm": 0.17268748581409454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39302932.0, "reward": 1.095313310623169, "reward_std": 0.47142454981803894, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6281249523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9187499284744263, "step": 3337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 844.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 547.625, "completions/mean_terminated_length": 547.625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.6676, "grad_norm": 0.06483941525220871, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39315422.0, "reward": 0.6706106662750244, "reward_std": 0.27574753761291504, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2787187397480011, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8418749570846558, "step": 3338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 491.6875, "completions/mean_terminated_length": 491.6875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.6678, "grad_norm": 0.13110174238681793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39327177.0, "reward": 1.0946099758148193, "reward_std": 0.2972029447555542, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.8767187595367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4468749761581421, "step": 3339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 664.0, "completions/max_terminated_length": 664.0, "completions/mean_length": 487.0625, "completions/mean_terminated_length": 487.0625, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.668, "grad_norm": 0.1248399093747139, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39338602.0, "reward": 0.7421882152557373, "reward_std": 0.21063940227031708, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.484375, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.59375, "step": 3340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 381.0, "completions/max_terminated_length": 381.0, "completions/mean_length": 263.4375, "completions/mean_terminated_length": 263.4375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.6682, "grad_norm": 0.21185539662837982, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39346545.0, "reward": 1.262160301208496, "reward_std": 0.4245700240135193, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.774318516254425, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9642499685287476, "step": 3341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 685.3125, "completions/mean_terminated_length": 662.7333374023438, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.6684, "grad_norm": 0.06242391839623451, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39361926.0, "reward": 0.5553855895996094, "reward_std": 0.2064216136932373, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17326875030994415, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8268750309944153, "step": 3342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 370.25, "completions/mean_terminated_length": 370.25, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.6686, "grad_norm": 0.11080527305603027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39371978.0, "reward": 1.315260648727417, "reward_std": 0.3965529203414917, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8180187344551086, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 3343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 290.5, "completions/mean_terminated_length": 290.5, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.6688, "grad_norm": 0.00015005693421699107, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39381010.0, "reward": 1.4988946914672852, "reward_std": 0.0003946831275243312, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977874755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 3344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 609.3125, "completions/mean_terminated_length": 609.3125, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.669, "grad_norm": 0.11688708513975143, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39395863.0, "reward": 0.8438165187835693, "reward_std": 0.456093430519104, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3751312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 752.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 553.6875, "completions/mean_terminated_length": 553.6875, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.6692, "grad_norm": 0.15190710127353668, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39408562.0, "reward": 1.0256102085113525, "reward_std": 0.48667752742767334, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5512187480926514, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 3346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 977.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 543.8125, "completions/mean_terminated_length": 543.8125, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.6694, "grad_norm": 0.26508858799934387, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39420943.0, "reward": 0.7180477380752563, "reward_std": 0.32431814074516296, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3110937476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 3347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 310.375, "completions/mean_terminated_length": 310.375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.6696, "grad_norm": 0.14844910800457, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39429621.0, "reward": 1.3802071809768677, "reward_std": 0.32456761598587036, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854125142097473, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 516.9375, "completions/mean_terminated_length": 516.9375, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.6698, "grad_norm": 0.002956958021968603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39441636.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 536.5, "completions/mean_terminated_length": 536.5, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.67, "grad_norm": 0.12570416927337646, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39455340.0, "reward": 0.7658600807189941, "reward_std": 0.30602434277534485, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.40671876072883606, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 3350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 959.8125, "completions/mean_terminated_length": 818.6000366210938, "completions/min_length": 681.0, "completions/min_terminated_length": 681.0, "epoch": 0.6702, "grad_norm": 0.0783606544137001, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39474953.0, "reward": 0.1989845484495163, "reward_std": 0.3591807782649994, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14796875417232513, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.18437500298023224, "step": 3351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 861.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 340.1875, "completions/mean_terminated_length": 340.1875, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.6704, "grad_norm": 0.1418340504169464, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39484540.0, "reward": 0.892812967300415, "reward_std": 0.17522911727428436, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7856249809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3187499940395355, "step": 3352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 418.0625, "completions/mean_terminated_length": 418.0625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.6706, "grad_norm": 0.0927138701081276, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39495869.0, "reward": 0.7225008010864258, "reward_std": 0.2038593292236328, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4449999928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 3353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 334.625, "completions/mean_terminated_length": 334.625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.6708, "grad_norm": 0.06317302584648132, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39505623.0, "reward": 1.4402105808258057, "reward_std": 0.23772314190864563, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9429187774658203, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 3354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 518.6875, "completions/mean_terminated_length": 518.6875, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.671, "grad_norm": 0.03471384197473526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39518322.0, "reward": 0.5973446369171143, "reward_std": 0.13293686509132385, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19468750059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 3355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 303.25, "completions/mean_terminated_length": 303.25, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.6712, "grad_norm": 0.12449733167886734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39526950.0, "reward": 0.6470510959625244, "reward_std": 0.3334040939807892, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16910000145435333, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "step": 3356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 324.0, "completions/max_terminated_length": 324.0, "completions/mean_length": 286.8125, "completions/mean_terminated_length": 286.8125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.6714, "grad_norm": 0.13137130439281464, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39535123.0, "reward": 1.439582109451294, "reward_std": 0.2375553995370865, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416625499725342, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 274.875, "completions/mean_terminated_length": 274.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.6716, "grad_norm": 0.09025359153747559, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39543425.0, "reward": 1.439638376235962, "reward_std": 0.23757043480873108, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9417749643325806, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 401.3125, "completions/mean_terminated_length": 401.3125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.6718, "grad_norm": 0.16585509479045868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39553158.0, "reward": 0.8539071083068848, "reward_std": 0.4491252899169922, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39531248807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 3359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 625.1875, "completions/mean_terminated_length": 492.25, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.672, "grad_norm": 0.1944267600774765, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39566921.0, "reward": 0.8271757960319519, "reward_std": 0.6475890278816223, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4668499827384949, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7175000309944153, "step": 3360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 408.25, "completions/mean_terminated_length": 408.25, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.6722, "grad_norm": 0.10269002616405487, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39577901.0, "reward": 1.4395915269851685, "reward_std": 0.23755794763565063, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416812658309937, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 3361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 518.5, "completions/mean_terminated_length": 518.5, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.6724, "grad_norm": 0.09972813725471497, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39590869.0, "reward": 1.0454696416854858, "reward_std": 0.4352804720401764, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6534374952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8250000476837158, "step": 3362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 438.0, "completions/mean_terminated_length": 438.0, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.6726, "grad_norm": 0.21656398475170135, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39601253.0, "reward": 1.0628671646118164, "reward_std": 0.5113912224769592, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5632312297821045, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.986875057220459, "step": 3363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 707.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 451.6875, "completions/mean_terminated_length": 451.6875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.6728, "grad_norm": 0.11461792141199112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39612480.0, "reward": 0.9950791001319885, "reward_std": 0.4665675759315491, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5526562929153442, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 3364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 303.0625, "completions/mean_terminated_length": 303.0625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.673, "grad_norm": 0.09735380858182907, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39620929.0, "reward": 1.3206915855407715, "reward_std": 0.38298988342285156, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8288812637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 846.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 550.8125, "completions/mean_terminated_length": 550.8125, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.6732, "grad_norm": 0.07065790891647339, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39633262.0, "reward": 0.6108603477478027, "reward_std": 0.23768793046474457, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15921874344348907, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 403.125, "completions/mean_terminated_length": 403.125, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.6734, "grad_norm": 0.2504579722881317, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39643696.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 310.5625, "completions/mean_terminated_length": 310.5625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.6736, "grad_norm": 0.10056914389133453, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39652025.0, "reward": 1.380601167678833, "reward_std": 0.3247215747833252, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.886199951171875, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 3368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 840.6875, "completions/mean_terminated_length": 779.5833740234375, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.6738, "grad_norm": 0.12331293523311615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39669428.0, "reward": 0.7017192840576172, "reward_std": 0.5277920365333557, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4659375250339508, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5375000238418579, "step": 3369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 786.4375, "completions/mean_terminated_length": 731.6154174804688, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.674, "grad_norm": 0.12747997045516968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39685835.0, "reward": 0.6451570391654968, "reward_std": 0.4594461917877197, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3528125286102295, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6187499761581421, "step": 3370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 374.625, "completions/mean_terminated_length": 374.625, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.6742, "grad_norm": 0.1513695865869522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39695381.0, "reward": 0.944041907787323, "reward_std": 0.5060054659843445, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45058128237724304, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 3371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 250.0, "completions/max_terminated_length": 250.0, "completions/mean_length": 182.4375, "completions/mean_terminated_length": 182.4375, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.6744, "grad_norm": 0.0006458863499574363, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39701916.0, "reward": 1.4996353387832642, "reward_std": 0.0005300732445903122, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9992687702178955, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 3372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 387.6875, "completions/mean_terminated_length": 387.6875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.6746, "grad_norm": 0.15088504552841187, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39711575.0, "reward": 1.3802915811538696, "reward_std": 0.3246006369590759, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855812549591064, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 494.625, "completions/mean_terminated_length": 459.3333435058594, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.6748, "grad_norm": 0.16220255196094513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39724065.0, "reward": 0.7501446008682251, "reward_std": 0.4664057493209839, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3127875030040741, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1001.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 663.0, "completions/mean_terminated_length": 663.0, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.675, "grad_norm": 0.1457800716161728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39738817.0, "reward": 0.9905165433883667, "reward_std": 0.4765179753303528, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5435312390327454, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 3375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 466.25, "completions/mean_terminated_length": 466.25, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.6752, "grad_norm": 0.11076424270868301, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39751157.0, "reward": 0.9892978072166443, "reward_std": 0.4409147799015045, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6035937666893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.778124988079071, "step": 3376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 321.0625, "completions/mean_terminated_length": 321.0625, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.6754, "grad_norm": 0.23499448597431183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39759894.0, "reward": 1.0237510204315186, "reward_std": 0.49057790637016296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 361.5625, "completions/mean_terminated_length": 361.5625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.6756, "grad_norm": 0.21465443074703217, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39769935.0, "reward": 1.163985013961792, "reward_std": 0.3907458484172821, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7029687166213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7406250238418579, "step": 3378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 526.4375, "completions/mean_terminated_length": 526.4375, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.6758, "grad_norm": 0.13455650210380554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39782070.0, "reward": 0.80445396900177, "reward_std": 0.3081035912036896, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.48390620946884155, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7593749761581421, "step": 3379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 274.0, "completions/max_terminated_length": 274.0, "completions/mean_length": 231.125, "completions/mean_terminated_length": 231.125, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.676, "grad_norm": 0.00045765601680614054, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39789400.0, "reward": 1.499100923538208, "reward_std": 0.000538655323907733, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998199999332428, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500667572021, "step": 3380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 343.3125, "completions/mean_terminated_length": 343.3125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.6762, "grad_norm": 0.20604056119918823, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39799133.0, "reward": 1.1642072200775146, "reward_std": 0.44994038343429565, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7034125328063965, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 3381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 683.3125, "completions/mean_terminated_length": 634.6428833007812, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.6764, "grad_norm": 0.07834908366203308, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39815282.0, "reward": 0.682331919670105, "reward_std": 0.41714122891426086, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3646624982357025, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7050000429153442, "step": 3382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 353.375, "completions/mean_terminated_length": 353.375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.6766, "grad_norm": 0.1746954470872879, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39825016.0, "reward": 0.8382037878036499, "reward_std": 0.28547418117523193, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5514062643051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5218750238418579, "step": 3383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 963.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 692.625, "completions/mean_terminated_length": 692.625, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.6768, "grad_norm": 0.06729215383529663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39841490.0, "reward": 0.6536728143692017, "reward_std": 0.2631607949733734, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2448437511920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 3384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 626.625, "completions/mean_terminated_length": 600.1333618164062, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.677, "grad_norm": 0.06106140464544296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39856332.0, "reward": 0.5979695320129395, "reward_std": 0.21019034087657928, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25843748450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 3385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 529.375, "completions/mean_terminated_length": 529.375, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.6772, "grad_norm": 0.10217802971601486, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39869762.0, "reward": 1.3181915283203125, "reward_std": 0.38735777139663696, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8238812685012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 3386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 379.0, "completions/max_terminated_length": 379.0, "completions/mean_length": 304.125, "completions/mean_terminated_length": 304.125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.6774, "grad_norm": 0.20990081131458282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39878692.0, "reward": 1.3189572095870972, "reward_std": 0.38696539402008057, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8254125118255615, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 452.8125, "completions/mean_terminated_length": 452.8125, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.6776, "grad_norm": 0.12275625765323639, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39889777.0, "reward": 0.7240603566169739, "reward_std": 0.38424602150917053, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26061874628067017, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 3388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 317.0, "completions/max_terminated_length": 317.0, "completions/mean_length": 228.9375, "completions/mean_terminated_length": 228.9375, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.6778, "grad_norm": 0.0005520209087990224, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39897008.0, "reward": 1.4994947910308838, "reward_std": 0.0005964276497252285, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989874958992004, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 3389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 941.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 617.375, "completions/mean_terminated_length": 617.375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.678, "grad_norm": 0.10693857818841934, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39910774.0, "reward": 0.8913285732269287, "reward_std": 0.23544976115226746, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.7201562523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.39687496423721313, "step": 3390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 433.1875, "completions/mean_terminated_length": 433.1875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.6782, "grad_norm": 0.11129220575094223, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39921545.0, "reward": 1.4366791248321533, "reward_std": 0.24978114664554596, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9358562231063843, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 452.4375, "completions/mean_terminated_length": 452.4375, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.6784, "grad_norm": 0.07799419015645981, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39933312.0, "reward": 1.3778914213180542, "reward_std": 0.2511104941368103, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8807812333106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.796875, "step": 3392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 289.0, "completions/mean_terminated_length": 289.0, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.6786, "grad_norm": 0.317251592874527, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39942016.0, "reward": 0.7863290905952454, "reward_std": 0.4249928593635559, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3226562440395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 445.9375, "completions/mean_terminated_length": 445.9375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.6788, "grad_norm": 0.12158078700304031, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39953759.0, "reward": 0.7550790309906006, "reward_std": 0.3756004869937897, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3226562440395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 3394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 452.0, "completions/max_terminated_length": 452.0, "completions/mean_length": 369.875, "completions/mean_terminated_length": 369.875, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.679, "grad_norm": 0.2013590931892395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39963997.0, "reward": 0.9673447608947754, "reward_std": 0.484319806098938, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4971875250339508, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 346.8125, "completions/mean_terminated_length": 346.8125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.6792, "grad_norm": 0.08448974043130875, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39973914.0, "reward": 1.4399946928024292, "reward_std": 0.23766563832759857, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424875378608704, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 3396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 354.3125, "completions/mean_terminated_length": 354.3125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.6794, "grad_norm": 0.16707244515419006, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39983551.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 625.4375, "completions/mean_terminated_length": 625.4375, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.6796, "grad_norm": 0.09112649410963058, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39997686.0, "reward": 1.439591646194458, "reward_std": 0.23755794763565063, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9416812658309937, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 3398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 246.0, "completions/max_terminated_length": 246.0, "completions/mean_length": 200.0625, "completions/mean_terminated_length": 200.0625, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.6798, "grad_norm": 0.00042173804831691086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40004279.0, "reward": 1.499657392501831, "reward_std": 0.000540734501555562, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9993125200271606, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "step": 3399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 385.125, "completions/mean_terminated_length": 385.125, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.68, "grad_norm": 0.07204395532608032, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40014217.0, "reward": 1.4375884532928467, "reward_std": 0.2473706156015396, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9376749992370605, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9724999666213989, "step": 3400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 350.125, "completions/mean_terminated_length": 350.125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.6802, "grad_norm": 0.3880288004875183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40024251.0, "reward": 0.962566614151001, "reward_std": 0.48842906951904297, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48763126134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 306.1875, "completions/mean_terminated_length": 306.1875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.6804, "grad_norm": 0.4954080283641815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40032654.0, "reward": 1.4052759408950806, "reward_std": 0.37474021315574646, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9355499744415283, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 809.25, "completions/mean_terminated_length": 759.6923217773438, "completions/min_length": 600.0, "completions/min_terminated_length": 600.0, "epoch": 0.6806, "grad_norm": 0.12142307311296463, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40049586.0, "reward": 0.5840629935264587, "reward_std": 0.3812778890132904, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4181250035762787, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.44999998807907104, "step": 3403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 736.5, "completions/mean_terminated_length": 717.3333740234375, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.6808, "grad_norm": 0.13996979594230652, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40065306.0, "reward": 0.8313288688659668, "reward_std": 0.4566464126110077, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4751562476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 792.0, "completions/mean_terminated_length": 758.857177734375, "completions/min_length": 581.0, "completions/min_terminated_length": 581.0, "epoch": 0.681, "grad_norm": 0.1074042096734047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40082762.0, "reward": 0.7926445603370667, "reward_std": 0.529312252998352, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39778751134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8112499713897705, "step": 3405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 463.4375, "completions/mean_terminated_length": 463.4375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.6812, "grad_norm": 0.22228172421455383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40094625.0, "reward": 0.8473446369171143, "reward_std": 0.41330191493034363, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.44468748569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 3406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 367.6875, "completions/mean_terminated_length": 367.6875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.6814, "grad_norm": 0.3870014250278473, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40104412.0, "reward": 0.8438165187835693, "reward_std": 0.456093430519104, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3751312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 482.1875, "completions/mean_terminated_length": 446.0666809082031, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.6816, "grad_norm": 0.3451702892780304, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40115743.0, "reward": 0.81421959400177, "reward_std": 0.49586209654808044, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37843751907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 3408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 402.6875, "completions/mean_terminated_length": 402.6875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.6818, "grad_norm": 0.13798387348651886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40125962.0, "reward": 1.4078915119171143, "reward_std": 0.26130160689353943, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9407812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 330.9375, "completions/mean_terminated_length": 330.9375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.682, "grad_norm": 0.1814812570810318, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40135065.0, "reward": 1.2682695388793945, "reward_std": 0.34012454748153687, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8490375280380249, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7787500023841858, "step": 3410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 350.8125, "completions/mean_terminated_length": 350.8125, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.6822, "grad_norm": 0.1585732251405716, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40144662.0, "reward": 1.034116506576538, "reward_std": 0.47950479388237, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5682312250137329, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9318749904632568, "step": 3411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 463.4375, "completions/mean_terminated_length": 463.4375, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.6824, "grad_norm": 0.07741431146860123, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40156541.0, "reward": 0.6108572483062744, "reward_std": 0.23970720171928406, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15921249985694885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 3412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 280.0, "completions/max_terminated_length": 280.0, "completions/mean_length": 238.4375, "completions/mean_terminated_length": 238.4375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.6826, "grad_norm": 0.00015140972391236573, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40163812.0, "reward": 1.4998385906219482, "reward_std": 0.0002974927774630487, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9996750354766846, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 3413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 637.75, "completions/mean_terminated_length": 612.0000610351562, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.6828, "grad_norm": 0.11159466952085495, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40178160.0, "reward": 0.6424416899681091, "reward_std": 0.38131147623062134, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22238123416900635, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8618749976158142, "step": 3414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 710.0, "completions/mean_length": 428.875, "completions/mean_terminated_length": 389.20001220703125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.683, "grad_norm": 0.1851169615983963, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40189230.0, "reward": 1.2299916744232178, "reward_std": 0.4981019198894501, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7724812626838684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 3415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 421.6875, "completions/mean_terminated_length": 421.6875, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.6832, "grad_norm": 0.20155592262744904, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40200793.0, "reward": 1.1406915187835693, "reward_std": 0.4774612784385681, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6563812494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 539.0625, "completions/mean_terminated_length": 469.7857360839844, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.6834, "grad_norm": 0.06503790616989136, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40214138.0, "reward": 0.5025008916854858, "reward_std": 0.22142644226551056, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12999999523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.793749988079071, "step": 3417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 398.0625, "completions/mean_terminated_length": 398.0625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.6836, "grad_norm": 0.3330991566181183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40224443.0, "reward": 0.9561386108398438, "reward_std": 0.4951012432575226, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4747749865055084, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 3418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 354.8125, "completions/mean_terminated_length": 354.8125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.6838, "grad_norm": 0.0001864010700955987, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40234312.0, "reward": 1.4989571571350098, "reward_std": 0.0004445533559191972, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979125261306763, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 3419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 644.875, "completions/mean_terminated_length": 644.875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.684, "grad_norm": 0.4079757630825043, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40248678.0, "reward": 1.0676414966583252, "reward_std": 0.5270522832870483, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6352812051773071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8731249570846558, "step": 3420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 707.8125, "completions/mean_terminated_length": 707.8125, "completions/min_length": 571.0, "completions/min_terminated_length": 571.0, "epoch": 0.6842, "grad_norm": 0.03604431450366974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40263891.0, "reward": 0.6492977142333984, "reward_std": 0.15398158133029938, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.29859375953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 3421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 504.625, "completions/mean_terminated_length": 470.0000305175781, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.6844, "grad_norm": 0.07549769431352615, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40276413.0, "reward": 1.2868914604187012, "reward_std": 0.4713890254497528, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8237812519073486, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9081250429153442, "step": 3422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 190.1875, "completions/mean_terminated_length": 190.1875, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.6846, "grad_norm": 0.0003266334824729711, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40283120.0, "reward": 1.499341607093811, "reward_std": 0.0006110992399044335, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986813068389893, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 3423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 439.375, "completions/mean_terminated_length": 439.375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.6848, "grad_norm": 0.08610127121210098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40295078.0, "reward": 0.6784385442733765, "reward_std": 0.32246747612953186, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2318750023841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 3424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 396.625, "completions/mean_terminated_length": 396.625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.685, "grad_norm": 0.15739849209785461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40306272.0, "reward": 0.9627853631973267, "reward_std": 0.4886854290962219, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48806875944137573, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 633.75, "completions/mean_terminated_length": 633.75, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.6852, "grad_norm": 0.13325117528438568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40320652.0, "reward": 0.8126165270805359, "reward_std": 0.4234830439090729, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3752312660217285, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 3426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 314.1875, "completions/mean_terminated_length": 314.1875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.6854, "grad_norm": 0.17781560122966766, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40330431.0, "reward": 0.8467853665351868, "reward_std": 0.45436131954193115, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38106876611709595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518750309944153, "step": 3427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 645.625, "completions/mean_terminated_length": 645.625, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.6856, "grad_norm": 0.05660362169146538, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40345849.0, "reward": 0.7795319557189941, "reward_std": 0.17436176538467407, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5590624809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5625, "step": 3428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 449.3125, "completions/mean_terminated_length": 449.3125, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.6858, "grad_norm": 0.045297153294086456, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40357086.0, "reward": 1.4676697254180908, "reward_std": 0.1260463148355484, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.997837483882904, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9087499976158142, "step": 3429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 710.0, "completions/max_terminated_length": 710.0, "completions/mean_length": 367.75, "completions/mean_terminated_length": 367.75, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.686, "grad_norm": 0.3571699261665344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40366970.0, "reward": 0.9189854264259338, "reward_std": 0.4690525233745575, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.46296873688697815, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 469.375, "completions/mean_terminated_length": 469.375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.6862, "grad_norm": 0.12685935199260712, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40378016.0, "reward": 0.8510944247245789, "reward_std": 0.32169219851493835, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5771874785423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6312499642372131, "step": 3431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1003.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 688.8125, "completions/mean_terminated_length": 688.8125, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.6864, "grad_norm": 0.1736866682767868, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40392637.0, "reward": 1.126950979232788, "reward_std": 0.44670143723487854, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6913999915122986, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 3432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 640.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 575.375, "completions/mean_terminated_length": 575.375, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.6866, "grad_norm": 0.1431429535150528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40406115.0, "reward": 1.0833477973937988, "reward_std": 0.4430193305015564, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6666936874389648, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 3433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 446.6875, "completions/mean_terminated_length": 446.6875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.6868, "grad_norm": 0.1188577339053154, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40417134.0, "reward": 1.216094732284546, "reward_std": 0.30822521448135376, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.9321874380111694, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5750000476837158, "step": 3434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 384.4375, "completions/mean_terminated_length": 384.4375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.687, "grad_norm": 0.3062479496002197, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40428821.0, "reward": 0.6212073564529419, "reward_std": 0.36826983094215393, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1799125075340271, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 3435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 530.9375, "completions/mean_terminated_length": 498.0666809082031, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.6872, "grad_norm": 0.19291190803050995, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40443796.0, "reward": 0.9876415729522705, "reward_std": 0.5356926321983337, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5377812385559082, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8731249570846558, "step": 3436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 622.9375, "completions/mean_terminated_length": 596.2000122070312, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.6874, "grad_norm": 0.08568210154771805, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40457651.0, "reward": 1.1897602081298828, "reward_std": 0.4374110698699951, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.8795187473297119, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6081249713897705, "step": 3437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 769.125, "completions/mean_terminated_length": 710.3077392578125, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.6876, "grad_norm": 0.18176373839378357, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40473829.0, "reward": 0.7726568579673767, "reward_std": 0.4937762916088104, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5453125238418579, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.518750011920929, "step": 3438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 458.5, "completions/mean_terminated_length": 458.5, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.6878, "grad_norm": 0.18152306973934174, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40485213.0, "reward": 1.2585729360580444, "reward_std": 0.4306614398956299, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.767143726348877, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 3439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 251.0, "completions/max_terminated_length": 251.0, "completions/mean_length": 198.875, "completions/mean_terminated_length": 198.875, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.688, "grad_norm": 0.0004394409479573369, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40492315.0, "reward": 1.4995794296264648, "reward_std": 0.0005786861875094473, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991562366485596, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 3440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 551.875, "completions/mean_terminated_length": 551.875, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.6882, "grad_norm": 0.06250996887683868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40504665.0, "reward": 1.4392197132110596, "reward_std": 0.23746047914028168, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9409375190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 668.0625, "completions/mean_terminated_length": 644.3333740234375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.6884, "grad_norm": 0.09771332144737244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40519978.0, "reward": 0.7543601989746094, "reward_std": 0.4278472661972046, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3837187588214874, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7856249809265137, "step": 3442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 862.0, "completions/mean_length": 750.125, "completions/mean_terminated_length": 731.86669921875, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.6886, "grad_norm": 0.1483708918094635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40536684.0, "reward": 0.758672297000885, "reward_std": 0.3222057819366455, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6423437595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4156250059604645, "step": 3443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 241.5625, "completions/mean_terminated_length": 241.5625, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.6888, "grad_norm": 0.22108231484889984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40544277.0, "reward": 1.3188042640686035, "reward_std": 0.3885582983493805, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8251062631607056, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 3444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1001.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 619.5625, "completions/mean_terminated_length": 619.5625, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.689, "grad_norm": 0.07522467523813248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40558222.0, "reward": 0.844844400882721, "reward_std": 0.2322896271944046, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6271874904632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.581250011920929, "step": 3445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 527.25, "completions/mean_terminated_length": 494.13336181640625, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.6892, "grad_norm": 0.1830495148897171, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40570882.0, "reward": 1.3137383460998535, "reward_std": 0.4362680912017822, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8774750232696533, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8362499475479126, "step": 3446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 258.0, "completions/max_terminated_length": 258.0, "completions/mean_length": 169.5, "completions/mean_terminated_length": 169.5, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 0.6894, "grad_norm": 0.0006578756729140878, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40577466.0, "reward": 1.499351143836975, "reward_std": 0.0006215283647179604, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987000226974487, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 3447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 419.5625, "completions/mean_terminated_length": 419.5625, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.6896, "grad_norm": 0.16831612586975098, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40588883.0, "reward": 1.3431260585784912, "reward_std": 0.347194641828537, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8737499713897705, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 3448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 418.375, "completions/mean_terminated_length": 418.375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.6898, "grad_norm": 0.1157725527882576, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40599177.0, "reward": 1.3783321380615234, "reward_std": 0.32952845096588135, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8816624879837036, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 226.125, "completions/mean_terminated_length": 226.125, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.69, "grad_norm": 0.17728127539157867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40606555.0, "reward": 1.4399542808532715, "reward_std": 0.237654909491539, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424062967300415, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 3450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 838.375, "completions/mean_terminated_length": 776.5, "completions/min_length": 521.0, "completions/min_terminated_length": 521.0, "epoch": 0.6902, "grad_norm": 0.1705365628004074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40624033.0, "reward": 0.5518755912780762, "reward_std": 0.3713737726211548, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.35374999046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.48124998807907104, "step": 3451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 997.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 624.5625, "completions/mean_terminated_length": 624.5625, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.6904, "grad_norm": 0.12665876746177673, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40638234.0, "reward": 0.5552322864532471, "reward_std": 0.18719474971294403, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1729625016450882, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 3452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 484.1875, "completions/mean_terminated_length": 448.20001220703125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.6906, "grad_norm": 0.21856647729873657, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40649645.0, "reward": 0.9903509616851807, "reward_std": 0.5420709252357483, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5432000160217285, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 3453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 383.5625, "completions/mean_terminated_length": 340.86669921875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.6908, "grad_norm": 0.3502325415611267, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40659414.0, "reward": 1.0210943222045898, "reward_std": 0.4587126672267914, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6046874523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6499999761581421, "step": 3454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 628.1875, "completions/mean_terminated_length": 628.1875, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.691, "grad_norm": 0.0792047530412674, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40673273.0, "reward": 0.6525166630744934, "reward_std": 0.3311721086502075, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18003100156784058, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9679999947547913, "step": 3455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 466.375, "completions/mean_terminated_length": 466.375, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.6912, "grad_norm": 0.18360060453414917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40685071.0, "reward": 0.853829026222229, "reward_std": 0.45022690296173096, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39515626430511475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 3456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 531.25, "completions/mean_terminated_length": 531.25, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.6914, "grad_norm": 0.19648195803165436, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40697507.0, "reward": 1.1790322065353394, "reward_std": 0.4570697546005249, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6705625057220459, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 3457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 577.75, "completions/mean_terminated_length": 577.75, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.6916, "grad_norm": 0.18859536945819855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40710927.0, "reward": 0.9922664761543274, "reward_std": 0.473943829536438, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5470312833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 3458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 336.875, "completions/mean_terminated_length": 336.875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.6918, "grad_norm": 0.09081560373306274, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40720029.0, "reward": 1.4396729469299316, "reward_std": 0.23757977783679962, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9418437480926514, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 3459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 630.6875, "completions/mean_terminated_length": 604.4666748046875, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.692, "grad_norm": 0.0821341946721077, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40735080.0, "reward": 0.6807040572166443, "reward_std": 0.28036218881607056, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.42390623688697815, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5968749523162842, "step": 3460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 669.25, "completions/mean_terminated_length": 587.3846435546875, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.6922, "grad_norm": 0.09624332934617996, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40750076.0, "reward": 0.6782817244529724, "reward_std": 0.3766724765300751, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5440624952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.36250001192092896, "step": 3461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 423.8125, "completions/mean_terminated_length": 423.8125, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.6924, "grad_norm": 0.16332823038101196, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40760969.0, "reward": 1.0829696655273438, "reward_std": 0.4865889549255371, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6034375429153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 336.4375, "completions/mean_terminated_length": 336.4375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.6926, "grad_norm": 0.18068856000900269, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40770512.0, "reward": 0.6926259994506836, "reward_std": 0.33516445755958557, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2602500021457672, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 3463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 358.4375, "completions/mean_terminated_length": 358.4375, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.6928, "grad_norm": 0.0695963054895401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40779943.0, "reward": 0.6433602571487427, "reward_std": 0.12948811054229736, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2867187261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "step": 3464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 469.8125, "completions/mean_terminated_length": 469.8125, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.693, "grad_norm": 0.15823239088058472, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40792404.0, "reward": 1.2017947435379028, "reward_std": 0.4539577066898346, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7160874605178833, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 3465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 453.125, "completions/mean_terminated_length": 453.125, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.6932, "grad_norm": 0.16036993265151978, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40804390.0, "reward": 1.0486476421356201, "reward_std": 0.39715659618377686, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7222937345504761, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6818749904632568, "step": 3466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 326.0, "completions/max_terminated_length": 326.0, "completions/mean_length": 256.875, "completions/mean_terminated_length": 256.875, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.6934, "grad_norm": 0.09627307951450348, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40812564.0, "reward": 1.4318509101867676, "reward_std": 0.23751606047153473, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9262000322341919, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287499785423279, "step": 3467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 484.5625, "completions/mean_terminated_length": 484.5625, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.6936, "grad_norm": 0.13884283602237701, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40824365.0, "reward": 1.400816559791565, "reward_std": 0.2754529118537903, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9266312122344971, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9356250166893005, "step": 3468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 510.75, "completions/mean_terminated_length": 476.5333557128906, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.6938, "grad_norm": 0.235244482755661, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40836425.0, "reward": 1.1927226781845093, "reward_std": 0.5656298398971558, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7604438066482544, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 3469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 334.8125, "completions/mean_terminated_length": 334.8125, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.694, "grad_norm": 0.14246679842472076, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40845558.0, "reward": 1.1971728801727295, "reward_std": 0.4624685049057007, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7068437337875366, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 3470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 440.0, "completions/mean_terminated_length": 440.0, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.6942, "grad_norm": 0.006318175233900547, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40856838.0, "reward": 0.5336228013038635, "reward_std": 0.018026450648903847, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06724374741315842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 3471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 483.0, "completions/max_terminated_length": 483.0, "completions/mean_length": 392.5, "completions/mean_terminated_length": 392.5, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.6944, "grad_norm": 0.13225364685058594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40867694.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 273.3125, "completions/mean_terminated_length": 273.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.6946, "grad_norm": 0.000352351285982877, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40876499.0, "reward": 1.4991633892059326, "reward_std": 0.0005518344114534557, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998324990272522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 3473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 384.75, "completions/mean_terminated_length": 384.75, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.6948, "grad_norm": 0.18525363504886627, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40886383.0, "reward": 1.1425665616989136, "reward_std": 0.4750525653362274, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6601312756538391, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 307.0, "completions/max_terminated_length": 307.0, "completions/mean_length": 225.375, "completions/mean_terminated_length": 225.375, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.695, "grad_norm": 0.21993371844291687, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40893669.0, "reward": 0.7832041382789612, "reward_std": 0.42683979868888855, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31640625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 574.25, "completions/mean_terminated_length": 544.2667236328125, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.6952, "grad_norm": 0.13415072858333588, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40906377.0, "reward": 0.6874884366989136, "reward_std": 0.4240095317363739, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24997499585151672, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 368.375, "completions/mean_terminated_length": 368.375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.6954, "grad_norm": 0.02590738795697689, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40916159.0, "reward": 0.5628761053085327, "reward_std": 0.07530082762241364, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12575000524520874, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 3477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 429.4375, "completions/mean_terminated_length": 429.4375, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.6956, "grad_norm": 0.2810351252555847, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40926966.0, "reward": 0.8828135132789612, "reward_std": 0.4496086537837982, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.390625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 3478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 312.5, "completions/mean_terminated_length": 312.5, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.6958, "grad_norm": 0.1883556991815567, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40937006.0, "reward": 1.319035291671753, "reward_std": 0.3870042562484741, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8255687355995178, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 3479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 535.6875, "completions/mean_terminated_length": 535.6875, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.696, "grad_norm": 0.22935692965984344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40949929.0, "reward": 1.0407822132110596, "reward_std": 0.43904349207878113, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6440625190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 3480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 632.625, "completions/mean_terminated_length": 576.7142944335938, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.6962, "grad_norm": 0.16612260043621063, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40964083.0, "reward": 0.7799633741378784, "reward_std": 0.5336697101593018, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37242498993873596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8337500095367432, "step": 3481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 569.375, "completions/mean_terminated_length": 569.375, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.6964, "grad_norm": 0.1516997218132019, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40976937.0, "reward": 1.1935131549835205, "reward_std": 0.34008416533470154, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.699524998664856, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6462500095367432, "step": 3482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 309.0, "completions/max_terminated_length": 309.0, "completions/mean_length": 269.5625, "completions/mean_terminated_length": 269.5625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.6966, "grad_norm": 0.15803758800029755, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40986594.0, "reward": 1.2574197053909302, "reward_std": 0.4327603578567505, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7648375034332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 3483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 595.375, "completions/mean_terminated_length": 595.375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.6968, "grad_norm": 0.08533231168985367, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40999816.0, "reward": 0.684532105922699, "reward_std": 0.26892200112342834, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.30656248331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 3484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 449.6875, "completions/mean_terminated_length": 449.6875, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.697, "grad_norm": 0.0002098055265378207, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41010995.0, "reward": 1.499313473701477, "reward_std": 0.0005835460615344346, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986250400543213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "step": 3485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 612.9375, "completions/mean_terminated_length": 554.2142944335938, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.6972, "grad_norm": 0.10318293422460556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41024818.0, "reward": 0.5272667407989502, "reward_std": 0.2599702775478363, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1795312464237213, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 390.875, "completions/mean_terminated_length": 390.875, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.6974, "grad_norm": 0.16671748459339142, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41035504.0, "reward": 1.3803603649139404, "reward_std": 0.32462751865386963, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8857187628746033, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 692.6875, "completions/mean_terminated_length": 645.357177734375, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.6976, "grad_norm": 0.2276885211467743, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41050795.0, "reward": 1.0773441791534424, "reward_std": 0.4661254584789276, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.5296875238418579, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.375, "step": 3488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 524.0625, "completions/mean_terminated_length": 524.0625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.6978, "grad_norm": 0.061680492013692856, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41062940.0, "reward": 0.7901568412780762, "reward_std": 0.17697015404701233, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5803124904632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5625, "step": 3489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 357.6875, "completions/mean_terminated_length": 357.6875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.698, "grad_norm": 0.19712673127651215, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41072775.0, "reward": 1.1425009965896606, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 603.0625, "completions/mean_terminated_length": 575.0, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.6982, "grad_norm": 0.08292189240455627, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41086472.0, "reward": 1.2832633256912231, "reward_std": 0.4403269290924072, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8790249824523926, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7849999666213989, "step": 3491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 496.625, "completions/mean_terminated_length": 496.625, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.6984, "grad_norm": 0.11934184283018112, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41098402.0, "reward": 1.4058916568756104, "reward_std": 0.26585817337036133, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9367812871932983, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256249666213989, "step": 3492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 645.625, "completions/mean_terminated_length": 620.4000244140625, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.6986, "grad_norm": 0.17648887634277344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41112908.0, "reward": 0.8953289985656738, "reward_std": 0.5158336758613586, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4781562387943268, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 3493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 295.75, "completions/mean_terminated_length": 295.75, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.6988, "grad_norm": 0.1186976283788681, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41121288.0, "reward": 1.4397978782653809, "reward_std": 0.23761312663555145, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9420937895774841, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 430.0625, "completions/mean_terminated_length": 430.0625, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.699, "grad_norm": 0.10128700733184814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41132345.0, "reward": 1.3802071809768677, "reward_std": 0.324567586183548, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854125142097473, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 296.3125, "completions/mean_terminated_length": 296.3125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.6992, "grad_norm": 0.12700849771499634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41141134.0, "reward": 0.5880917906761169, "reward_std": 0.24372915923595428, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11368125677108765, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 3496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 585.3125, "completions/mean_terminated_length": 585.3125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.6994, "grad_norm": 0.12508682906627655, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41155571.0, "reward": 0.7166414260864258, "reward_std": 0.38025087118148804, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3707812428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 3497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1008.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 753.625, "completions/mean_terminated_length": 753.625, "completions/min_length": 563.0, "completions/min_terminated_length": 563.0, "epoch": 0.6996, "grad_norm": 0.06931047141551971, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41171741.0, "reward": 0.8223444819450378, "reward_std": 0.29812705516815186, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5196874737739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 3498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 322.75, "completions/mean_terminated_length": 322.75, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.6998, "grad_norm": 0.1005459800362587, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41180793.0, "reward": 1.438194751739502, "reward_std": 0.24490562081336975, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9388875365257263, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 3499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 604.5, "completions/mean_terminated_length": 507.69232177734375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.7, "grad_norm": 0.18759283423423767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41194081.0, "reward": 0.8193756341934204, "reward_std": 0.5582239627838135, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5137500166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.612500011920929, "step": 3500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 411.9375, "completions/mean_terminated_length": 324.5, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.7002, "grad_norm": 0.16425290703773499, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41204656.0, "reward": 0.6041290163993835, "reward_std": 0.3983120024204254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20825624465942383, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8268749713897705, "step": 3501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 909.0, "completions/mean_terminated_length": 717.3333740234375, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.7004, "grad_norm": 0.09590893983840942, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41223456.0, "reward": 0.346640944480896, "reward_std": 0.5314134955406189, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19328126311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.31562501192092896, "step": 3502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 432.25, "completions/mean_terminated_length": 432.25, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.7006, "grad_norm": 0.028754938393831253, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41234852.0, "reward": 0.585141658782959, "reward_std": 0.09153030067682266, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17028124630451202, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9043749570846558, "step": 3503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 370.625, "completions/mean_terminated_length": 370.625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.7008, "grad_norm": 0.11869283765554428, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41244382.0, "reward": 0.6966953277587891, "reward_std": 0.3988589644432068, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20588749647140503, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9900000095367432, "step": 3504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 374.125, "completions/mean_terminated_length": 374.125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.701, "grad_norm": 0.16626039147377014, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41254384.0, "reward": 1.197291612625122, "reward_std": 0.46255117654800415, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7070812582969666, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 380.0, "completions/mean_terminated_length": 380.0, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.7012, "grad_norm": 0.06964185833930969, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41264064.0, "reward": 1.4400135278701782, "reward_std": 0.2376706898212433, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9425250291824341, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 3506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 858.375, "completions/mean_terminated_length": 803.1666870117188, "completions/min_length": 527.0, "completions/min_terminated_length": 527.0, "epoch": 0.7014, "grad_norm": 0.11936992406845093, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41282246.0, "reward": 0.49171939492225647, "reward_std": 0.3292328417301178, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23343749344348907, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.581250011920929, "step": 3507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 451.3125, "completions/mean_terminated_length": 451.3125, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.7016, "grad_norm": 0.15563823282718658, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41293963.0, "reward": 0.81156325340271, "reward_std": 0.27356263995170593, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5606249570846558, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.574999988079071, "step": 3508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 301.375, "completions/mean_terminated_length": 301.375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.7018, "grad_norm": 0.14350280165672302, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41302769.0, "reward": 0.9025759100914001, "reward_std": 0.47703468799591064, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43015000224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 649.8125, "completions/mean_terminated_length": 624.86669921875, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.702, "grad_norm": 0.08425013720989227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41317550.0, "reward": 0.6467822790145874, "reward_std": 0.37840139865875244, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23106250166893005, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8574999570846558, "step": 3510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 467.9375, "completions/mean_terminated_length": 430.86669921875, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.7022, "grad_norm": 0.2166198045015335, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41329197.0, "reward": 1.0744383335113525, "reward_std": 0.5295966863632202, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6488749980926514, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8450000286102295, "step": 3511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 770.25, "completions/mean_terminated_length": 516.5, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.7024, "grad_norm": 0.20586897432804108, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41345873.0, "reward": 0.4817971885204315, "reward_std": 0.5370027422904968, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2760937809944153, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.19687500596046448, "step": 3512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 341.8125, "completions/mean_terminated_length": 341.8125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.7026, "grad_norm": 0.35727059841156006, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41355358.0, "reward": 1.1841853857040405, "reward_std": 0.4665670692920685, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6808687448501587, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9443750381469727, "step": 3513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 395.125, "completions/mean_terminated_length": 395.125, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.7028, "grad_norm": 0.2500569522380829, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41365472.0, "reward": 0.9152353405952454, "reward_std": 0.4684962332248688, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4554687440395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 549.0625, "completions/mean_terminated_length": 517.4000244140625, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.703, "grad_norm": 0.16924616694450378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41378497.0, "reward": 0.8980475664138794, "reward_std": 0.431163489818573, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6085937023162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 3515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 438.5625, "completions/mean_terminated_length": 438.5625, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.7032, "grad_norm": 0.05651610717177391, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41390154.0, "reward": 0.5962135195732117, "reward_std": 0.15835151076316833, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1924249827861786, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 3516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 255.75, "completions/mean_terminated_length": 255.75, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.7034, "grad_norm": 0.22875264286994934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41397974.0, "reward": 1.377944827079773, "reward_std": 0.3313332796096802, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.880887508392334, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 3517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 676.0, "completions/max_terminated_length": 676.0, "completions/mean_length": 433.6875, "completions/mean_terminated_length": 433.6875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.7036, "grad_norm": 0.38764896988868713, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41409073.0, "reward": 1.1655290126800537, "reward_std": 0.5263014435768127, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7060562372207642, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.90562504529953, "step": 3518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 428.5625, "completions/mean_terminated_length": 428.5625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.7038, "grad_norm": 0.17221234738826752, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41419802.0, "reward": 1.348435401916504, "reward_std": 0.3363792896270752, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8843687772750854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 3519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 761.6875, "completions/mean_terminated_length": 701.1538696289062, "completions/min_length": 537.0, "completions/min_terminated_length": 537.0, "epoch": 0.704, "grad_norm": 0.043762583285570145, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41436021.0, "reward": 0.4313414692878723, "reward_std": 0.21461521089076996, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05018124729394913, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7868750095367432, "step": 3520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 367.4375, "completions/mean_terminated_length": 367.4375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.7042, "grad_norm": 0.19241495430469513, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41446108.0, "reward": 1.3802103996276855, "reward_std": 0.32456883788108826, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854187726974487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 3521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 571.375, "completions/mean_terminated_length": 506.71429443359375, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.7044, "grad_norm": 0.18148626387119293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41458946.0, "reward": 0.7070320844650269, "reward_std": 0.46439212560653687, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2890625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.737500011920929, "step": 3522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 440.0625, "completions/mean_terminated_length": 440.0625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.7046, "grad_norm": 0.17538589239120483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41469987.0, "reward": 0.7501571178436279, "reward_std": 0.38009175658226013, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3128125071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 3523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 915.1875, "completions/mean_terminated_length": 806.375, "completions/min_length": 624.0, "completions/min_terminated_length": 624.0, "epoch": 0.7048, "grad_norm": 0.14825254678726196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41489814.0, "reward": 0.2998442053794861, "reward_std": 0.3163924813270569, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09968749433755875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4437499940395355, "step": 3524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 438.0625, "completions/mean_terminated_length": 438.0625, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.705, "grad_norm": 0.16667765378952026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41501559.0, "reward": 0.950782060623169, "reward_std": 0.4331054985523224, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.526562511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 3525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 410.9375, "completions/mean_terminated_length": 410.9375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.7052, "grad_norm": 0.07828603684902191, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41511574.0, "reward": 0.619991660118103, "reward_std": 0.24022836983203888, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1774812489748001, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9356249570846558, "step": 3526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 642.5625, "completions/mean_terminated_length": 642.5625, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.7054, "grad_norm": 0.09928019344806671, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41525807.0, "reward": 1.382969856262207, "reward_std": 0.316731333732605, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8909374475479126, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 353.6875, "completions/mean_terminated_length": 353.6875, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.7056, "grad_norm": 0.07215601205825806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41535274.0, "reward": 1.4394510984420776, "reward_std": 0.2375228852033615, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413999915122986, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 3528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 624.125, "completions/mean_terminated_length": 624.125, "completions/min_length": 505.0, "completions/min_terminated_length": 505.0, "epoch": 0.7058, "grad_norm": 0.13873150944709778, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41549884.0, "reward": 1.0195915699005127, "reward_std": 0.49514785408973694, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5391812324523926, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 3529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 319.0, "completions/max_terminated_length": 319.0, "completions/mean_length": 269.3125, "completions/mean_terminated_length": 269.3125, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.706, "grad_norm": 0.12753652036190033, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41558321.0, "reward": 1.4373729228973389, "reward_std": 0.24731312692165375, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9372438192367554, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 3530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 413.6875, "completions/mean_terminated_length": 413.6875, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.7062, "grad_norm": 0.1399586945772171, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41568636.0, "reward": 0.6755478382110596, "reward_std": 0.3234710097312927, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22609375417232513, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 3531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 537.625, "completions/mean_terminated_length": 537.625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.7064, "grad_norm": 0.10858805477619171, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41581046.0, "reward": 0.7369539141654968, "reward_std": 0.19398510456085205, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4739062786102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.659375011920929, "step": 3532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 389.125, "completions/mean_terminated_length": 389.125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.7066, "grad_norm": 0.12094443291425705, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41591128.0, "reward": 1.3764512538909912, "reward_std": 0.3363298773765564, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8779000043869019, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 3533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 239.0, "completions/max_terminated_length": 239.0, "completions/mean_length": 202.9375, "completions/mean_terminated_length": 202.9375, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.7068, "grad_norm": 0.11826933175325394, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41598007.0, "reward": 1.4403479099273682, "reward_std": 0.23775959014892578, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.943193793296814, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 3534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 248.4375, "completions/mean_terminated_length": 248.4375, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.707, "grad_norm": 0.3352540135383606, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41605518.0, "reward": 1.102007269859314, "reward_std": 0.47197625041007996, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6415125131607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9337500333786011, "step": 3535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 451.25, "completions/mean_terminated_length": 451.25, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.7072, "grad_norm": 0.14281587302684784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41617442.0, "reward": 1.040541410446167, "reward_std": 0.36491185426712036, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7685812711715698, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5393750071525574, "step": 3536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 555.75, "completions/mean_terminated_length": 524.5333862304688, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.7074, "grad_norm": 0.13725091516971588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41630542.0, "reward": 0.7210947275161743, "reward_std": 0.4201718270778656, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31718748807907104, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 3537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 634.1875, "completions/mean_terminated_length": 608.2000122070312, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.7076, "grad_norm": 0.1749432235956192, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41644561.0, "reward": 1.221379041671753, "reward_std": 0.44789186120033264, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8177562355995178, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.753125011920929, "step": 3538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 635.25, "completions/mean_terminated_length": 458.54547119140625, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.7078, "grad_norm": 0.4382662773132324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41658293.0, "reward": 0.7628908157348633, "reward_std": 0.5630274415016174, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.2757812440395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.203125, "step": 3539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 641.6875, "completions/mean_terminated_length": 553.4615478515625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.708, "grad_norm": 0.09667064249515533, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41672560.0, "reward": 0.6638286113739014, "reward_std": 0.4179355502128601, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.39015626907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.47187498211860657, "step": 3540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 543.375, "completions/mean_terminated_length": 511.3333740234375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.7082, "grad_norm": 0.10952619463205338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41684998.0, "reward": 0.8875789642333984, "reward_std": 0.4761717915534973, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5251562595367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.753125011920929, "step": 3541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 794.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 508.3125, "completions/mean_terminated_length": 508.3125, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.7084, "grad_norm": 0.022780071943998337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41696987.0, "reward": 0.5613293051719666, "reward_std": 0.06011839583516121, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12265624850988388, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 420.25, "completions/mean_terminated_length": 420.25, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.7086, "grad_norm": 0.13182970881462097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41707519.0, "reward": 0.7356727719306946, "reward_std": 0.3817944824695587, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2838437557220459, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 3543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 745.5, "completions/mean_terminated_length": 745.5, "completions/min_length": 485.0, "completions/min_terminated_length": 485.0, "epoch": 0.7088, "grad_norm": 0.08232270926237106, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41725367.0, "reward": 1.1671757698059082, "reward_std": 0.3452244997024536, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.8343499898910522, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7725000381469727, "step": 3544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 360.5, "completions/mean_terminated_length": 360.5, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.709, "grad_norm": 0.1582518368959427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41736447.0, "reward": 1.4398199319839478, "reward_std": 0.23761914670467377, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421374797821045, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 3545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 548.75, "completions/mean_terminated_length": 548.75, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.7092, "grad_norm": 0.21148085594177246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41749323.0, "reward": 0.8483601808547974, "reward_std": 0.3698081076145172, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3842187523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.778124988079071, "step": 3546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 423.875, "completions/mean_terminated_length": 423.875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.7094, "grad_norm": 0.01182539016008377, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41760249.0, "reward": 0.5496821999549866, "reward_std": 0.02202288620173931, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09936249256134033, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 278.0625, "completions/mean_terminated_length": 278.0625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.7096, "grad_norm": 0.17713257670402527, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41768602.0, "reward": 1.3206915855407715, "reward_std": 0.38298988342285156, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8288812637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 693.9375, "completions/mean_terminated_length": 617.7692260742188, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.7098, "grad_norm": 0.14102600514888763, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41783801.0, "reward": 0.687734842300415, "reward_std": 0.42394933104515076, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6254687309265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.24062500894069672, "step": 3549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 549.75, "completions/mean_terminated_length": 482.0000305175781, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.71, "grad_norm": 0.11790498346090317, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41796437.0, "reward": 0.6705350279808044, "reward_std": 0.3390366733074188, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46606874465942383, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.47062498331069946, "step": 3550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 452.1875, "completions/mean_terminated_length": 452.1875, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.7102, "grad_norm": 0.006206403486430645, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41807576.0, "reward": 0.5414759516716003, "reward_std": 0.015640748664736748, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08294999599456787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 874.5, "completions/mean_terminated_length": 824.6666870117188, "completions/min_length": 469.0, "completions/min_terminated_length": 469.0, "epoch": 0.7104, "grad_norm": 0.13351020216941833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41827136.0, "reward": 0.6592190861701965, "reward_std": 0.4088231325149536, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5684375166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3125, "step": 3552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 346.5, "completions/mean_terminated_length": 346.5, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.7106, "grad_norm": 0.10709146410226822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41837368.0, "reward": 1.3751980066299438, "reward_std": 0.33779841661453247, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8753937482833862, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 348.8125, "completions/mean_terminated_length": 348.8125, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.7108, "grad_norm": 0.18129613995552063, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41846677.0, "reward": 1.0733357667922974, "reward_std": 0.49950215220451355, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5841687917709351, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 3554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 790.625, "completions/mean_terminated_length": 650.6000366210938, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.711, "grad_norm": 0.09082643687725067, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41863407.0, "reward": 0.41984423995018005, "reward_std": 0.36680033802986145, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2146874964237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.4625000059604645, "step": 3555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 501.375, "completions/mean_terminated_length": 501.375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.7112, "grad_norm": 0.13642443716526031, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41875157.0, "reward": 1.09195077419281, "reward_std": 0.420817494392395, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.683899998664856, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.815000057220459, "step": 3556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 445.8125, "completions/mean_terminated_length": 445.8125, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.7114, "grad_norm": 0.00927971862256527, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41886514.0, "reward": 0.5484386086463928, "reward_std": 0.024534961208701134, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09687499701976776, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 376.75, "completions/mean_terminated_length": 376.75, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.7116, "grad_norm": 0.12280627340078354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41896046.0, "reward": 0.6309133768081665, "reward_std": 0.24079637229442596, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19932499527931213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9225000143051147, "step": 3558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 438.25, "completions/mean_terminated_length": 354.5714416503906, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.7118, "grad_norm": 0.1638612598180771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41906754.0, "reward": 1.1333602666854858, "reward_std": 0.581024706363678, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7042187452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 3559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 400.625, "completions/mean_terminated_length": 400.625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.712, "grad_norm": 0.21095694601535797, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41917004.0, "reward": 1.0060919523239136, "reward_std": 0.5090393424034119, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.512181282043457, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 3560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 481.0, "completions/mean_terminated_length": 481.0, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.7122, "grad_norm": 0.09650446474552155, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41928524.0, "reward": 0.6407039165496826, "reward_std": 0.15416617691516876, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2814062535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8093750476837158, "step": 3561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 595.0, "completions/mean_terminated_length": 595.0, "completions/min_length": 449.0, "completions/min_terminated_length": 449.0, "epoch": 0.7124, "grad_norm": 0.10531773418188095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41942028.0, "reward": 1.3785417079925537, "reward_std": 0.3296104073524475, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8820812702178955, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 998.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 630.1875, "completions/mean_terminated_length": 630.1875, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.7126, "grad_norm": 0.1340724229812622, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41956175.0, "reward": 0.9706259369850159, "reward_std": 0.4409501552581787, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5037500262260437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 3563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 577.0, "completions/max_terminated_length": 577.0, "completions/mean_length": 495.6875, "completions/mean_terminated_length": 495.6875, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.7128, "grad_norm": 0.2414623647928238, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41968282.0, "reward": 0.7692071199417114, "reward_std": 0.402451753616333, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28841251134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 3564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 710.375, "completions/mean_terminated_length": 638.0, "completions/min_length": 493.0, "completions/min_terminated_length": 493.0, "epoch": 0.713, "grad_norm": 0.15186884999275208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41983712.0, "reward": 0.8146101236343384, "reward_std": 0.5903348326683044, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.441718727350235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7531249523162842, "step": 3565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 647.25, "completions/mean_terminated_length": 560.3077392578125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.7132, "grad_norm": 0.14204862713813782, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41998340.0, "reward": 0.8294538855552673, "reward_std": 0.553242564201355, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5339062213897705, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5531250238418579, "step": 3566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 339.5625, "completions/mean_terminated_length": 339.5625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.7134, "grad_norm": 0.1829574704170227, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42007677.0, "reward": 1.147500991821289, "reward_std": 0.4699709713459015, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6700000166893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 3567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 319.375, "completions/mean_terminated_length": 319.375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.7136, "grad_norm": 0.1883677840232849, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42016643.0, "reward": 1.2023260593414307, "reward_std": 0.455091655254364, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7171499729156494, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 3568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1004.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 694.8125, "completions/mean_terminated_length": 694.8125, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.7138, "grad_norm": 0.12546712160110474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42031744.0, "reward": 1.3155040740966797, "reward_std": 0.3404831290245056, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8810062408447266, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8668750524520874, "step": 3569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 759.0, "completions/mean_terminated_length": 721.1428833007812, "completions/min_length": 560.0, "completions/min_terminated_length": 560.0, "epoch": 0.714, "grad_norm": 0.05571402609348297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42048256.0, "reward": 0.5435791015625, "reward_std": 0.27386030554771423, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2121562361717224, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6981250047683716, "step": 3570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 388.0625, "completions/mean_terminated_length": 388.0625, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.7142, "grad_norm": 0.13474991917610168, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42058081.0, "reward": 0.7523415088653564, "reward_std": 0.3868640959262848, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.317181259393692, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 3571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 547.75, "completions/mean_terminated_length": 547.75, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.7144, "grad_norm": 0.05790786072611809, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42070861.0, "reward": 1.060469150543213, "reward_std": 0.22112222015857697, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8709374666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.375, "step": 3572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 301.0, "completions/mean_terminated_length": 301.0, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.7146, "grad_norm": 0.15512040257453918, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42079405.0, "reward": 0.6523949503898621, "reward_std": 0.3312416970729828, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1797875016927719, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 3573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 831.5, "completions/mean_terminated_length": 804.0000610351562, "completions/min_length": 678.0, "completions/min_terminated_length": 678.0, "epoch": 0.7148, "grad_norm": 0.12923499941825867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42096757.0, "reward": 1.1020821332931519, "reward_std": 0.5736572742462158, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7041624784469604, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7875000238418579, "step": 3574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 698.5625, "completions/mean_terminated_length": 623.4615478515625, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.715, "grad_norm": 0.08910021930932999, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42112606.0, "reward": 0.5910038948059082, "reward_std": 0.4385037124156952, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24450625479221344, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7143750190734863, "step": 3575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 301.875, "completions/mean_terminated_length": 301.875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.7152, "grad_norm": 0.1387525349855423, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42121564.0, "reward": 0.655935525894165, "reward_std": 0.3296279013156891, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18686875700950623, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 274.5, "completions/mean_terminated_length": 274.5, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.7154, "grad_norm": 0.11288714408874512, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42129716.0, "reward": 1.4379104375839233, "reward_std": 0.24745620787143707, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9383187294006348, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 3577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 617.8125, "completions/mean_terminated_length": 590.7333374023438, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.7156, "grad_norm": 0.11093800514936447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42145169.0, "reward": 1.0171570777893066, "reward_std": 0.5035824179649353, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5968124866485596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7400000095367432, "step": 3578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 572.25, "completions/mean_terminated_length": 507.71429443359375, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.7158, "grad_norm": 0.14273019134998322, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42158517.0, "reward": 1.2542164325714111, "reward_std": 0.5130065679550171, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8209312558174133, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7118749618530273, "step": 3579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 487.6875, "completions/mean_terminated_length": 487.6875, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.716, "grad_norm": 0.12167935073375702, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42170304.0, "reward": 0.848516583442688, "reward_std": 0.4529082477092743, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38453125953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 451.375, "completions/mean_terminated_length": 451.375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.7162, "grad_norm": 0.06479766964912415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42182390.0, "reward": 0.6367853879928589, "reward_std": 0.18490897119045258, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2735687494277954, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7831250429153442, "step": 3581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 602.75, "completions/mean_terminated_length": 602.75, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.7164, "grad_norm": 0.07348956912755966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42197458.0, "reward": 0.6675009727478027, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20999999344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 799.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 476.4375, "completions/mean_terminated_length": 476.4375, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.7166, "grad_norm": 0.1543298065662384, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42210441.0, "reward": 1.2153915166854858, "reward_std": 0.43573567271232605, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7432812452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 480.5625, "completions/mean_terminated_length": 480.5625, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.7168, "grad_norm": 0.12643946707248688, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42222130.0, "reward": 0.9532820582389832, "reward_std": 0.45477095246315, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5315625071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 3584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 475.5625, "completions/mean_terminated_length": 475.5625, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.717, "grad_norm": 0.16173610091209412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42233643.0, "reward": 1.0441415309906006, "reward_std": 0.4760846793651581, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5882812738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 3585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 426.25, "completions/mean_terminated_length": 426.25, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.7172, "grad_norm": 0.12421239167451859, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42245647.0, "reward": 0.7808228135108948, "reward_std": 0.4282363951206207, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3116437494754791, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 490.6875, "completions/mean_terminated_length": 490.6875, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.7174, "grad_norm": 0.09367115795612335, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42258586.0, "reward": 0.6675759553909302, "reward_std": 0.3246918022632599, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21014998853206635, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 496.25, "completions/mean_terminated_length": 496.25, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.7176, "grad_norm": 0.13087400794029236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42270654.0, "reward": 0.9694538712501526, "reward_std": 0.41072383522987366, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6264062523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 3588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 491.4375, "completions/mean_terminated_length": 455.933349609375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.7178, "grad_norm": 0.0780007466673851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42283205.0, "reward": 0.8103129863739014, "reward_std": 0.275088369846344, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6831250190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3812499940395355, "step": 3589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 329.0625, "completions/mean_terminated_length": 329.0625, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.718, "grad_norm": 0.14824919402599335, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42292486.0, "reward": 1.2019416093826294, "reward_std": 0.45482364296913147, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.716381311416626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 614.8125, "completions/mean_terminated_length": 614.8125, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.7182, "grad_norm": 0.05294887349009514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42307459.0, "reward": 0.6335948705673218, "reward_std": 0.171444833278656, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2671875059604645, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8125, "step": 3591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 499.0, "completions/max_terminated_length": 499.0, "completions/mean_length": 362.625, "completions/mean_terminated_length": 362.625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.7184, "grad_norm": 0.05508536472916603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42317885.0, "reward": 0.9409347772598267, "reward_std": 0.12441953271627426, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8818687200546265, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.2306250035762787, "step": 3592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 360.0, "completions/mean_terminated_length": 360.0, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.7186, "grad_norm": 0.15148663520812988, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42326989.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 344.8125, "completions/mean_terminated_length": 344.8125, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.7188, "grad_norm": 0.00027137642609886825, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42336986.0, "reward": 1.499313473701477, "reward_std": 0.000583555200137198, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986250400543213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9675000309944153, "step": 3594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 570.625, "completions/mean_terminated_length": 505.857177734375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.719, "grad_norm": 0.09094829112291336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42350820.0, "reward": 0.44694146513938904, "reward_std": 0.22218851745128632, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08138124644756317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7706249952316284, "step": 3595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 445.5, "completions/mean_terminated_length": 445.5, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.7192, "grad_norm": 0.17780998349189758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42362204.0, "reward": 1.0189697742462158, "reward_std": 0.4958362877368927, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5379375219345093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 927.8125, "completions/mean_terminated_length": 831.625, "completions/min_length": 664.0, "completions/min_terminated_length": 664.0, "epoch": 0.7194, "grad_norm": 0.17132189869880676, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42381353.0, "reward": 0.33375048637390137, "reward_std": 0.4140787124633789, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10500000417232513, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4749999940395355, "step": 3597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 424.0, "completions/mean_terminated_length": 424.0, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.7196, "grad_norm": 0.09417029470205307, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42391689.0, "reward": 1.4400978088378906, "reward_std": 0.2376931607723236, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9426937699317932, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 3598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 462.9375, "completions/mean_terminated_length": 462.9375, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.7198, "grad_norm": 0.12795378267765045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42403032.0, "reward": 1.228191614151001, "reward_std": 0.4280479848384857, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7688812017440796, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081249833106995, "step": 3599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 923.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 653.125, "completions/mean_terminated_length": 653.125, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.72, "grad_norm": 0.09927544742822647, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42419082.0, "reward": 0.6213164329528809, "reward_std": 0.22780796885490417, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3051312565803528, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7956249713897705, "step": 3600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 619.9375, "completions/mean_terminated_length": 593.0000610351562, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.7202, "grad_norm": 0.10466060042381287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42432745.0, "reward": 0.7427352666854858, "reward_std": 0.4239806532859802, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3604687452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8093749284744263, "step": 3601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 282.0625, "completions/mean_terminated_length": 282.0625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.7204, "grad_norm": 0.0663590356707573, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42441194.0, "reward": 1.4398760795593262, "reward_std": 0.23763403296470642, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422500133514404, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 3602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 624.6875, "completions/mean_terminated_length": 567.6428833007812, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.7206, "grad_norm": 0.1513068526983261, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42455365.0, "reward": 0.6694538593292236, "reward_std": 0.49087613821029663, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.33890625834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 363.0, "completions/mean_terminated_length": 363.0, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.7208, "grad_norm": 0.00014808042033109814, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42464917.0, "reward": 1.4989041090011597, "reward_std": 0.0004187046142760664, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9978062510490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 348.1875, "completions/mean_terminated_length": 348.1875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.721, "grad_norm": 0.21986453235149384, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42474808.0, "reward": 1.2001978158950806, "reward_std": 0.45768997073173523, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7128937244415283, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 321.625, "completions/mean_terminated_length": 321.625, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.7212, "grad_norm": 0.2575456500053406, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42484642.0, "reward": 1.024438500404358, "reward_std": 0.4905736744403839, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5488749742507935, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 3606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 904.3125, "completions/mean_terminated_length": 750.4285888671875, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.7214, "grad_norm": 0.07433151453733444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42502903.0, "reward": 0.3023441433906555, "reward_std": 0.41978776454925537, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10468749701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.4124999940395355, "step": 3607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 774.0, "completions/mean_terminated_length": 716.3077392578125, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.7216, "grad_norm": 0.12334571033716202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42519271.0, "reward": 0.6945322751998901, "reward_std": 0.5350270867347717, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3265624940395355, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 3608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 642.375, "completions/mean_terminated_length": 616.933349609375, "completions/min_length": 463.0, "completions/min_terminated_length": 463.0, "epoch": 0.7218, "grad_norm": 0.1340377777814865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42533805.0, "reward": 0.8342697024345398, "reward_std": 0.4969483017921448, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4185374975204468, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8562500476837158, "step": 3609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 239.0, "completions/max_terminated_length": 239.0, "completions/mean_length": 167.5625, "completions/mean_terminated_length": 167.5625, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.722, "grad_norm": 3.515607750159688e-05, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42539926.0, "reward": 1.49995756149292, "reward_std": 4.813246414414607e-05, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9999125003814697, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9925000071525574, "step": 3610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 442.3125, "completions/mean_terminated_length": 442.3125, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.7222, "grad_norm": 0.09196027368307114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42550907.0, "reward": 0.6532041430473328, "reward_std": 0.2633630633354187, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24390624463558197, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.871874988079071, "step": 3611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 384.125, "completions/mean_terminated_length": 384.125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.7224, "grad_norm": 0.061675164848566055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42561645.0, "reward": 0.9371098279953003, "reward_std": 0.15326792001724243, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.874218761920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.20937500894069672, "step": 3612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 544.8125, "completions/mean_terminated_length": 544.8125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.7226, "grad_norm": 0.0712757259607315, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42575066.0, "reward": 0.7051106691360474, "reward_std": 0.39436572790145874, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2227187603712082, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 3613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 232.0625, "completions/mean_terminated_length": 232.0625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.7228, "grad_norm": 0.1822671890258789, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42582555.0, "reward": 1.2458539009094238, "reward_std": 0.30510199069976807, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8042062520980835, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6206250190734863, "step": 3614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 647.125, "completions/mean_terminated_length": 622.0000610351562, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.723, "grad_norm": 0.12461747974157333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42596637.0, "reward": 0.9361726641654968, "reward_std": 0.5035456418991089, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6223437786102295, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.590624988079071, "step": 3615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 813.5625, "completions/mean_terminated_length": 687.2999877929688, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.7232, "grad_norm": 0.11894556134939194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42613286.0, "reward": 0.6191287040710449, "reward_std": 0.6093674302101135, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3632562458515167, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5331249833106995, "step": 3616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 419.4375, "completions/mean_terminated_length": 419.4375, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.7234, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42623853.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1005.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 494.125, "completions/mean_terminated_length": 494.125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.7236, "grad_norm": 0.008602303452789783, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42635631.0, "reward": 0.5443665981292725, "reward_std": 0.02057151310145855, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08873124420642853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 551.6875, "completions/mean_terminated_length": 551.6875, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.7238, "grad_norm": 0.031062165275216103, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42648650.0, "reward": 0.611641526222229, "reward_std": 0.09892504662275314, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22328123450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.871874988079071, "step": 3619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 371.0, "completions/max_terminated_length": 371.0, "completions/mean_length": 319.875, "completions/mean_terminated_length": 319.875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.724, "grad_norm": 1.1438704859756399e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42658216.0, "reward": 1.4999668598175049, "reward_std": 2.4354189008590765e-05, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9999312162399292, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9931250214576721, "step": 3620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 520.125, "completions/mean_terminated_length": 520.125, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.7242, "grad_norm": 0.15943101048469543, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42670042.0, "reward": 1.0799980163574219, "reward_std": 0.49099215865135193, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5974937677383423, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 288.0, "completions/mean_terminated_length": 288.0, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.7244, "grad_norm": 0.1520136594772339, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42677994.0, "reward": 0.8013946413993835, "reward_std": 0.4213591516017914, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35278749465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9300000071525574, "step": 3622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 333.125, "completions/mean_terminated_length": 333.125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.7246, "grad_norm": 0.13964074850082397, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42688188.0, "reward": 0.6566010117530823, "reward_std": 0.3291696608066559, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1881999969482422, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 3623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 318.6875, "completions/mean_terminated_length": 318.6875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.7248, "grad_norm": 0.0002495352237019688, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42697623.0, "reward": 1.499100923538208, "reward_std": 0.00053863984066993, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998199999332428, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500667572021, "step": 3624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 552.0, "completions/mean_terminated_length": 520.5333862304688, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.725, "grad_norm": 0.11141494661569595, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42711207.0, "reward": 0.6842069029808044, "reward_std": 0.36564311385154724, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.43091249465942383, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6581249833106995, "step": 3625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 525.6875, "completions/mean_terminated_length": 492.4667053222656, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.7252, "grad_norm": 0.15448938310146332, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42724098.0, "reward": 1.114675760269165, "reward_std": 0.46335265040397644, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7918499708175659, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6675000190734863, "step": 3626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 453.4375, "completions/mean_terminated_length": 453.4375, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.7254, "grad_norm": 0.15996688604354858, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42735241.0, "reward": 1.2662508487701416, "reward_std": 0.36926451325416565, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8450000286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 3627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 539.9375, "completions/mean_terminated_length": 507.66668701171875, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.7256, "grad_norm": 0.14074279367923737, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42747800.0, "reward": 1.3732696771621704, "reward_std": 0.3871364891529083, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9340375065803528, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8675000071525574, "step": 3628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 412.125, "completions/mean_terminated_length": 412.125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.7258, "grad_norm": 0.07596957683563232, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42758202.0, "reward": 0.8781223893165588, "reward_std": 0.1973622888326645, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7562437653541565, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3368750214576721, "step": 3629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 775.9375, "completions/mean_terminated_length": 583.0, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.726, "grad_norm": 0.15900243818759918, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42774889.0, "reward": 0.391484797000885, "reward_std": 0.4692583382129669, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22046874463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4312499761581421, "step": 3630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 366.5, "completions/mean_terminated_length": 366.5, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.7262, "grad_norm": 0.22886459529399872, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42784401.0, "reward": 0.9064571857452393, "reward_std": 0.4747054874897003, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4379124939441681, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9500000476837158, "step": 3631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 620.6875, "completions/mean_terminated_length": 527.6154174804688, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.7264, "grad_norm": 0.18593011796474457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42798492.0, "reward": 0.8302101492881775, "reward_std": 0.5919675230979919, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4729187488555908, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7256250381469727, "step": 3632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 522.375, "completions/mean_terminated_length": 488.933349609375, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.7266, "grad_norm": 0.11720389127731323, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42812770.0, "reward": 0.6899977922439575, "reward_std": 0.42390546202659607, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.254993736743927, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 3633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 388.8125, "completions/mean_terminated_length": 388.8125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.7268, "grad_norm": 0.15821239352226257, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42823375.0, "reward": 1.0239697694778442, "reward_std": 0.4908040463924408, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5479375123977661, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 745.625, "completions/mean_terminated_length": 681.3846435546875, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.727, "grad_norm": 0.0833943635225296, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42839737.0, "reward": 0.4748445451259613, "reward_std": 0.24495479464530945, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13718749582767487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.737500011920929, "step": 3635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 524.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 436.625, "completions/mean_terminated_length": 436.625, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.7272, "grad_norm": 0.13995034992694855, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42850467.0, "reward": 0.726250946521759, "reward_std": 0.38161933422088623, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26499998569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 3636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 485.125, "completions/mean_terminated_length": 485.125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.7274, "grad_norm": 0.10434883087873459, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42861669.0, "reward": 0.749532163143158, "reward_std": 0.22077809274196625, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4990624785423279, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6000000238418579, "step": 3637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 511.3125, "completions/mean_terminated_length": 511.3125, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.7276, "grad_norm": 0.17916753888130188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42874794.0, "reward": 1.1576321125030518, "reward_std": 0.44947949051856995, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.690262496471405, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8725000023841858, "step": 3638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 448.8125, "completions/mean_terminated_length": 448.8125, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.7278, "grad_norm": 0.167479008436203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42886327.0, "reward": 0.9644509553909302, "reward_std": 0.4868180751800537, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49140000343322754, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 680.8125, "completions/mean_terminated_length": 657.933349609375, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.728, "grad_norm": 0.1246592104434967, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42901460.0, "reward": 0.8656383752822876, "reward_std": 0.5167403221130371, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4187749922275543, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8712500333786011, "step": 3640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 413.75, "completions/mean_terminated_length": 413.75, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.7282, "grad_norm": 0.1674671173095703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42911952.0, "reward": 1.1689321994781494, "reward_std": 0.4532569646835327, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.712862491607666, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9100000262260437, "step": 3641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 258.0, "completions/max_terminated_length": 258.0, "completions/mean_length": 197.4375, "completions/mean_terminated_length": 197.4375, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.7284, "grad_norm": 0.3757787346839905, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42918631.0, "reward": 1.3134729862213135, "reward_std": 0.400309681892395, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8144437074661255, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 3642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 839.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 664.75, "completions/mean_terminated_length": 664.75, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.7286, "grad_norm": 0.10418067127466202, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42933235.0, "reward": 0.6777353286743164, "reward_std": 0.33797234296798706, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35546875, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.684374988079071, "step": 3643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 560.5625, "completions/mean_terminated_length": 529.6666870117188, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.7288, "grad_norm": 0.11679564416408539, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42945932.0, "reward": 0.9870319366455078, "reward_std": 0.46808090806007385, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6615625023841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 3644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 638.5625, "completions/mean_terminated_length": 583.5, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.729, "grad_norm": 0.1050940752029419, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42960469.0, "reward": 0.623907208442688, "reward_std": 0.41047927737236023, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24781249463558197, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 3645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 598.625, "completions/mean_terminated_length": 598.625, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.7292, "grad_norm": 0.12398727238178253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42974351.0, "reward": 0.9589697122573853, "reward_std": 0.4919963479042053, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4804375171661377, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 447.0625, "completions/mean_terminated_length": 447.0625, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.7294, "grad_norm": 0.08421123772859573, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42985744.0, "reward": 1.3833822011947632, "reward_std": 0.3168926239013672, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8917624950408936, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 3647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 807.0625, "completions/mean_terminated_length": 757.0, "completions/min_length": 590.0, "completions/min_terminated_length": 590.0, "epoch": 0.7296, "grad_norm": 0.24866996705532074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43002929.0, "reward": 0.9414569735527039, "reward_std": 0.6201401352882385, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5704125165939331, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.75, "step": 3648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 584.0, "completions/max_terminated_length": 584.0, "completions/mean_length": 406.125, "completions/mean_terminated_length": 406.125, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.7298, "grad_norm": 0.008619341067969799, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43014003.0, "reward": 0.5244948267936707, "reward_std": 0.019738739356398582, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0489874966442585, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 3649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 318.5625, "completions/mean_terminated_length": 318.5625, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.73, "grad_norm": 0.2796582877635956, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43022652.0, "reward": 1.0090231895446777, "reward_std": 0.5068986415863037, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5180437564849854, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750619888306, "step": 3650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 392.0, "completions/mean_terminated_length": 392.0, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.7302, "grad_norm": 0.0754910483956337, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43032812.0, "reward": 1.3800666332244873, "reward_std": 0.3245124816894531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 456.5625, "completions/mean_terminated_length": 456.5625, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.7304, "grad_norm": 0.161795973777771, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43044101.0, "reward": 1.049725890159607, "reward_std": 0.47717955708503723, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5994499921798706, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 3652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 522.9375, "completions/mean_terminated_length": 522.9375, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.7306, "grad_norm": 0.11801216751337051, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43056564.0, "reward": 1.2106852531433105, "reward_std": 0.4428795874118805, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7338687181472778, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9443750381469727, "step": 3653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 315.3125, "completions/mean_terminated_length": 315.3125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.7308, "grad_norm": 0.21012872457504272, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43065641.0, "reward": 0.7201353311538696, "reward_std": 0.38714200258255005, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25276875495910645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 288.9375, "completions/mean_terminated_length": 288.9375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.731, "grad_norm": 0.07763636857271194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43074232.0, "reward": 1.4391133785247803, "reward_std": 0.2374342978000641, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9407249689102173, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 288.3125, "completions/mean_terminated_length": 288.3125, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.7312, "grad_norm": 0.13779668509960175, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43082925.0, "reward": 1.055859923362732, "reward_std": 0.22970286011695862, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8617187738418579, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.25312501192092896, "step": 3656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 321.6875, "completions/mean_terminated_length": 321.6875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.7314, "grad_norm": 0.07571028172969818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43091912.0, "reward": 1.4392197132110596, "reward_std": 0.23746049404144287, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9409375190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 422.75, "completions/mean_terminated_length": 422.75, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.7316, "grad_norm": 0.1006397157907486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43102788.0, "reward": 1.0516413450241089, "reward_std": 0.29398247599601746, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8532812595367432, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.37812501192092896, "step": 3658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 576.6875, "completions/mean_terminated_length": 546.86669921875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.7318, "grad_norm": 0.07271679490804672, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43115839.0, "reward": 0.631641685962677, "reward_std": 0.3024336099624634, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2632812261581421, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7406250238418579, "step": 3659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 394.8125, "completions/mean_terminated_length": 394.8125, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.732, "grad_norm": 0.19607101380825043, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43125500.0, "reward": 0.7375009059906006, "reward_std": 0.38003286719322205, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2874999940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 3660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 435.75, "completions/mean_terminated_length": 435.75, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.7322, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43136920.0, "reward": 0.5487509369850159, "reward_std": 0.0, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09749999642372131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 313.0, "completions/max_terminated_length": 313.0, "completions/mean_length": 250.0, "completions/mean_terminated_length": 250.0, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.7324, "grad_norm": 0.2173762321472168, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43144712.0, "reward": 1.4377822875976562, "reward_std": 0.24479541182518005, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9380625486373901, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 3662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 824.6875, "completions/mean_terminated_length": 734.0909423828125, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.7326, "grad_norm": 0.1362951099872589, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43162643.0, "reward": 0.5833600163459778, "reward_std": 0.5413249135017395, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2917187809944153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6031249761581421, "step": 3663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 369.25, "completions/mean_terminated_length": 369.25, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.7328, "grad_norm": 0.18867246806621552, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43172135.0, "reward": 0.9006103873252869, "reward_std": 0.4789556562900543, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42621874809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 908.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 470.125, "completions/mean_terminated_length": 470.125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.733, "grad_norm": 0.021611513569951057, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43183785.0, "reward": 0.5554259419441223, "reward_std": 0.07229045033454895, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11084999889135361, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9387500286102295, "step": 3665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 367.9375, "completions/mean_terminated_length": 367.9375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.7332, "grad_norm": 0.15351073443889618, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43193320.0, "reward": 1.2640602588653564, "reward_std": 0.41976240277290344, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7781187295913696, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 3666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 371.5625, "completions/mean_terminated_length": 371.5625, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.7334, "grad_norm": 0.23234939575195312, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43204081.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 553.5, "completions/mean_terminated_length": 522.1333618164062, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.7336, "grad_norm": 0.11191517114639282, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43219097.0, "reward": 0.7943754196166992, "reward_std": 0.2846721112728119, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6512500047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4125000238418579, "step": 3668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 596.5, "completions/mean_terminated_length": 568.0, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.7338, "grad_norm": 0.13901300728321075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43233633.0, "reward": 1.0439850091934204, "reward_std": 0.4684285819530487, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5879687070846558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6531250476837158, "step": 3669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 617.9375, "completions/mean_terminated_length": 559.9285888671875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.734, "grad_norm": 0.1017170250415802, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43247120.0, "reward": 0.611172616481781, "reward_std": 0.3340153396129608, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2848437428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7406249642372131, "step": 3670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 801.0625, "completions/mean_terminated_length": 769.2142944335938, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.7342, "grad_norm": 0.0788579136133194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43264225.0, "reward": 0.6917974352836609, "reward_std": 0.30502811074256897, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5085937976837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5093750357627869, "step": 3671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 340.4375, "completions/mean_terminated_length": 294.8666687011719, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.7344, "grad_norm": 0.2797202467918396, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43273608.0, "reward": 0.9441165328025818, "reward_std": 0.5289958119392395, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5132312774658203, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8756250143051147, "step": 3672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 448.0, "completions/mean_terminated_length": 448.0, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.7346, "grad_norm": 0.19946028292179108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43284712.0, "reward": 0.7889727354049683, "reward_std": 0.42298299074172974, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32794374227523804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 3673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 523.1875, "completions/mean_terminated_length": 523.1875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.7348, "grad_norm": 0.12383471429347992, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43296715.0, "reward": 1.0275793075561523, "reward_std": 0.4901442527770996, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5551562309265137, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 288.0, "completions/max_terminated_length": 288.0, "completions/mean_length": 221.0, "completions/mean_terminated_length": 221.0, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.735, "grad_norm": 0.2535586655139923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43304043.0, "reward": 1.380441665649414, "reward_std": 0.3246592581272125, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.885881245136261, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 238.0, "completions/max_terminated_length": 238.0, "completions/mean_length": 165.375, "completions/mean_terminated_length": 165.375, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.7352, "grad_norm": 0.3423687517642975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43310465.0, "reward": 0.6335698962211609, "reward_std": 0.2731965482234955, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20463749766349792, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8887499570846558, "step": 3676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 893.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 613.1875, "completions/mean_terminated_length": 613.1875, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.7354, "grad_norm": 0.16288147866725922, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43324244.0, "reward": 0.991066575050354, "reward_std": 0.47341886162757874, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5446312427520752, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 3677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 782.125, "completions/mean_terminated_length": 726.3077392578125, "completions/min_length": 552.0, "completions/min_terminated_length": 552.0, "epoch": 0.7356, "grad_norm": 0.13582047820091248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43340918.0, "reward": 0.6365631222724915, "reward_std": 0.36813661456108093, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4606249928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4375, "step": 3678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 596.1875, "completions/mean_terminated_length": 567.6666870117188, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.7358, "grad_norm": 0.09814000129699707, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43355689.0, "reward": 0.6139727830886841, "reward_std": 0.3699492812156677, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16544374823570251, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 3679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 267.875, "completions/mean_terminated_length": 267.875, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.736, "grad_norm": 0.101066954433918, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43364727.0, "reward": 1.378029227256775, "reward_std": 0.33136630058288574, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8810563087463379, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96812504529953, "step": 3680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 967.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 621.6875, "completions/mean_terminated_length": 621.6875, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.7362, "grad_norm": 0.06365576386451721, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43378706.0, "reward": 0.8086723685264587, "reward_std": 0.2053699791431427, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6173437237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5031249523162842, "step": 3681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 418.8125, "completions/mean_terminated_length": 418.8125, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.7364, "grad_norm": 0.23706132173538208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43389071.0, "reward": 0.7918758392333984, "reward_std": 0.36279985308647156, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3962499797344208, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.862500011920929, "step": 3682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 354.0, "completions/max_terminated_length": 354.0, "completions/mean_length": 243.625, "completions/mean_terminated_length": 243.625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.7366, "grad_norm": 0.1975240409374237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43396745.0, "reward": 1.4350792169570923, "reward_std": 0.2493620216846466, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9326562285423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 3683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 607.75, "completions/mean_terminated_length": 607.75, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.7368, "grad_norm": 0.1643195003271103, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43410565.0, "reward": 1.3182384967803955, "reward_std": 0.38997066020965576, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8239749670028687, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 3684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 436.0, "completions/max_terminated_length": 436.0, "completions/mean_length": 325.8125, "completions/mean_terminated_length": 325.8125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.737, "grad_norm": 0.21477669477462769, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43419826.0, "reward": 1.0214009284973145, "reward_std": 0.49324899911880493, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5428000092506409, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 3685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 596.6875, "completions/mean_terminated_length": 596.6875, "completions/min_length": 493.0, "completions/min_terminated_length": 493.0, "epoch": 0.7372, "grad_norm": 0.10478668659925461, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43433037.0, "reward": 1.3772417306900024, "reward_std": 0.3330438435077667, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8794811964035034, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 612.75, "completions/mean_terminated_length": 612.75, "completions/min_length": 435.0, "completions/min_terminated_length": 435.0, "epoch": 0.7374, "grad_norm": 0.018548576161265373, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43448825.0, "reward": 0.9642190933227539, "reward_std": 0.11110590398311615, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9284374713897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.16249999403953552, "step": 3687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 532.875, "completions/mean_terminated_length": 532.875, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.7376, "grad_norm": 0.09353125095367432, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43461527.0, "reward": 1.433516502380371, "reward_std": 0.22463591396808624, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9295312166213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 3688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 798.3125, "completions/mean_terminated_length": 723.0833740234375, "completions/min_length": 468.0, "completions/min_terminated_length": 468.0, "epoch": 0.7378, "grad_norm": 0.10528992861509323, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43477852.0, "reward": 0.5792850852012634, "reward_std": 0.46150192618370056, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28356873989105225, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6518750190734863, "step": 3689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 338.25, "completions/mean_terminated_length": 338.25, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.738, "grad_norm": 0.037821583449840546, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43487104.0, "reward": 0.5785884857177734, "reward_std": 0.11730210483074188, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1571750044822693, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 3690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 430.5625, "completions/mean_terminated_length": 430.5625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.7382, "grad_norm": 0.17207609117031097, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43499481.0, "reward": 0.9971883296966553, "reward_std": 0.4209739565849304, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.49437499046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.737500011920929, "step": 3691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 352.0, "completions/mean_terminated_length": 352.0, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.7384, "grad_norm": 0.12866948544979095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43508841.0, "reward": 1.3208447694778442, "reward_std": 0.3830660879611969, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291875123977661, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 353.0, "completions/max_terminated_length": 353.0, "completions/mean_length": 294.5625, "completions/mean_terminated_length": 294.5625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.7386, "grad_norm": 0.23525744676589966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43517330.0, "reward": 0.9007666110992432, "reward_std": 0.4785188138484955, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42653125524520874, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 661.8125, "completions/mean_terminated_length": 661.8125, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.7388, "grad_norm": 0.1262367069721222, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43531823.0, "reward": 0.8168757557868958, "reward_std": 0.3190276622772217, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5087500214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7062499523162842, "step": 3694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 599.75, "completions/mean_terminated_length": 571.4666748046875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.739, "grad_norm": 0.16334131360054016, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43546059.0, "reward": 1.1331133842468262, "reward_std": 0.5206384658813477, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7037249803543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 3695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 456.3125, "completions/mean_terminated_length": 456.3125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.7392, "grad_norm": 0.17528153955936432, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43557184.0, "reward": 0.9543728828430176, "reward_std": 0.4966096580028534, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4712437391281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 3696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 577.625, "completions/mean_terminated_length": 577.625, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.7394, "grad_norm": 0.1418071836233139, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43570170.0, "reward": 1.0203633308410645, "reward_std": 0.4592282176017761, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6032249927520752, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 3697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 716.1875, "completions/mean_terminated_length": 613.5833740234375, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.7396, "grad_norm": 0.12827716767787933, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43585821.0, "reward": 0.6178913712501526, "reward_std": 0.5135882496833801, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2982812523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.671875, "step": 3698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 394.4375, "completions/mean_terminated_length": 394.4375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.7398, "grad_norm": 0.07947184890508652, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43596356.0, "reward": 1.4373822212219238, "reward_std": 0.2473156452178955, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9372625350952148, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 3699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 549.25, "completions/mean_terminated_length": 517.6000366210938, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.74, "grad_norm": 0.1769070327281952, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43609544.0, "reward": 0.9269416332244873, "reward_std": 0.5378391146659851, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47888123989105225, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 3700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 343.75, "completions/mean_terminated_length": 343.75, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.7402, "grad_norm": 0.0004263414302840829, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43618724.0, "reward": 1.4989728927612305, "reward_std": 0.0004791634273715317, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979437589645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 357.8125, "completions/mean_terminated_length": 357.8125, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.7404, "grad_norm": 0.07470376789569855, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43628305.0, "reward": 1.3784854412078857, "reward_std": 0.329588383436203, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8819687366485596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 493.8125, "completions/mean_terminated_length": 493.8125, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.7406, "grad_norm": 0.107888363301754, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43640734.0, "reward": 0.9207819700241089, "reward_std": 0.3881285488605499, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5915625095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7249999642372131, "step": 3703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 472.875, "completions/mean_terminated_length": 472.875, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.7408, "grad_norm": 0.3402720093727112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43652140.0, "reward": 0.9602727890014648, "reward_std": 0.4908653795719147, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48304376006126404, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 699.0625, "completions/mean_terminated_length": 652.6428833007812, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.741, "grad_norm": 0.1639719158411026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43667325.0, "reward": 1.0457789897918701, "reward_std": 0.5763442516326904, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.654056191444397, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7743750214576721, "step": 3705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 300.0, "completions/max_terminated_length": 300.0, "completions/mean_length": 249.3125, "completions/mean_terminated_length": 249.3125, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.7412, "grad_norm": 0.33752021193504333, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43675602.0, "reward": 1.0750010013580322, "reward_std": 0.49712586402893066, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5874999761581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 3706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 378.6875, "completions/mean_terminated_length": 378.6875, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.7414, "grad_norm": 0.007930930703878403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43686557.0, "reward": 0.5518384575843811, "reward_std": 0.028983699157834053, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1036750003695488, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 3707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 656.5, "completions/mean_terminated_length": 632.0000610351562, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.7416, "grad_norm": 0.16735094785690308, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43701205.0, "reward": 0.6037477254867554, "reward_std": 0.29139095544815063, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20749375224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 3708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 408.75, "completions/mean_terminated_length": 367.73333740234375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.7418, "grad_norm": 0.046385686844587326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43711153.0, "reward": 0.5502228736877441, "reward_std": 0.18649405241012573, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1629437506198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 3709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 829.125, "completions/mean_terminated_length": 816.1333618164062, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.742, "grad_norm": 0.1252945363521576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43728915.0, "reward": 0.7423442006111145, "reward_std": 0.3277386724948883, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6096875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4375, "step": 3710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 735.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 462.1875, "completions/mean_terminated_length": 462.1875, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.7422, "grad_norm": 0.17707665264606476, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43740086.0, "reward": 0.931628942489624, "reward_std": 0.4289402663707733, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5507562756538391, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7893750071525574, "step": 3711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 485.875, "completions/mean_terminated_length": 450.0000305175781, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.7424, "grad_norm": 0.27619609236717224, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43751684.0, "reward": 1.0053133964538574, "reward_std": 0.5312247276306152, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5731250047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.875, "step": 3712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 607.125, "completions/mean_terminated_length": 607.125, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.7426, "grad_norm": 0.14589212834835052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43765414.0, "reward": 1.075026035308838, "reward_std": 0.49712231755256653, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5875500440597534, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 3713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 416.125, "completions/mean_terminated_length": 416.125, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.7428, "grad_norm": 0.2064301073551178, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43775800.0, "reward": 1.1425665616989136, "reward_std": 0.4750525653362274, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6601312160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 438.0, "completions/max_terminated_length": 438.0, "completions/mean_length": 350.3125, "completions/mean_terminated_length": 350.3125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.743, "grad_norm": 0.2792944312095642, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43785805.0, "reward": 0.8456259369850159, "reward_std": 0.4547778367996216, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3787499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 472.125, "completions/mean_terminated_length": 472.125, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.7432, "grad_norm": 0.10540635138750076, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43797263.0, "reward": 0.8432009220123291, "reward_std": 0.45656633377075195, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37389999628067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 260.625, "completions/mean_terminated_length": 260.625, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.7434, "grad_norm": 0.287068635225296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43805049.0, "reward": 1.3210666179656982, "reward_std": 0.3831762671470642, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8296312689781189, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 3717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 664.8125, "completions/mean_terminated_length": 664.8125, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.7436, "grad_norm": 0.12300978600978851, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43819494.0, "reward": 1.1281540393829346, "reward_std": 0.4455357789993286, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6938062310218811, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8981250524520874, "step": 3718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 572.375, "completions/mean_terminated_length": 572.375, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.7438, "grad_norm": 0.10896222293376923, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43832588.0, "reward": 0.8475788831710815, "reward_std": 0.36859622597694397, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5076562762260437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.734375, "step": 3719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 680.0625, "completions/mean_terminated_length": 680.0625, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.744, "grad_norm": 0.087516188621521, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43848365.0, "reward": 0.7735943794250488, "reward_std": 0.19797322154045105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5471874475479126, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5687499642372131, "step": 3720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 444.875, "completions/mean_terminated_length": 444.875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.7442, "grad_norm": 0.07931984961032867, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43859451.0, "reward": 0.941797137260437, "reward_std": 0.13561560213565826, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8835937976837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.21562498807907104, "step": 3721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 993.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 657.6875, "completions/mean_terminated_length": 657.6875, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.7444, "grad_norm": 0.15628159046173096, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43875414.0, "reward": 0.9332040548324585, "reward_std": 0.4657438099384308, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49140626192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 3722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 442.0, "completions/max_terminated_length": 442.0, "completions/mean_length": 330.25, "completions/mean_terminated_length": 330.25, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.7446, "grad_norm": 0.18752093613147736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43884346.0, "reward": 1.3208227157592773, "reward_std": 0.3830551505088806, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829143762588501, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 432.0, "completions/max_terminated_length": 432.0, "completions/mean_length": 340.4375, "completions/mean_terminated_length": 340.4375, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.7448, "grad_norm": 0.12972263991832733, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43893361.0, "reward": 1.3779040575027466, "reward_std": 0.33131730556488037, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8808062076568604, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 3724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 495.1875, "completions/mean_terminated_length": 459.933349609375, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.745, "grad_norm": 0.09421142190694809, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43905604.0, "reward": 0.6963164806365967, "reward_std": 0.4209689795970917, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2676312327384949, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8831249475479126, "step": 3725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 749.6875, "completions/mean_terminated_length": 686.3846435546875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.7452, "grad_norm": 0.06408951431512833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43923055.0, "reward": 0.4458601474761963, "reward_std": 0.2212085723876953, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7718749642372131, "step": 3726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 257.0, "completions/mean_terminated_length": 257.0, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.7454, "grad_norm": 0.11762378364801407, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43930895.0, "reward": 1.4680603742599487, "reward_std": 0.1261504590511322, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9986187815666199, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9231250286102295, "step": 3727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 632.8125, "completions/mean_terminated_length": 606.7333374023438, "completions/min_length": 463.0, "completions/min_terminated_length": 463.0, "epoch": 0.7456, "grad_norm": 0.20231947302818298, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43945484.0, "reward": 0.7265633344650269, "reward_std": 0.4251004457473755, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.328125, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.824999988079071, "step": 3728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 183.5, "completions/mean_terminated_length": 183.5, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.7458, "grad_norm": 0.31007346510887146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43952004.0, "reward": 1.2594261169433594, "reward_std": 0.429636687040329, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7688499689102173, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 3729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 803.5, "completions/mean_terminated_length": 730.0, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.746, "grad_norm": 0.12240123003721237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43968828.0, "reward": 0.5651538372039795, "reward_std": 0.4624386429786682, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2553062438964844, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6556249856948853, "step": 3730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 285.6875, "completions/mean_terminated_length": 285.6875, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.7462, "grad_norm": 0.18493399024009705, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43977239.0, "reward": 1.3888134956359863, "reward_std": 0.30143991112709045, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9026249647140503, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512499570846558, "step": 3731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 513.5, "completions/mean_terminated_length": 513.5, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.7464, "grad_norm": 0.120947927236557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43989519.0, "reward": 1.2586698532104492, "reward_std": 0.42918661236763, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7673375010490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 481.6875, "completions/mean_terminated_length": 481.6875, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.7466, "grad_norm": 0.1001264750957489, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44001338.0, "reward": 0.7189791202545166, "reward_std": 0.38723334670066833, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2504562437534332, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 533.4375, "completions/mean_terminated_length": 533.4375, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.7468, "grad_norm": 0.0001542670652270317, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44014129.0, "reward": 1.4993228912353516, "reward_std": 0.0005930510233156383, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986437559127808, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 3734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 621.875, "completions/mean_terminated_length": 621.875, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.747, "grad_norm": 0.03921344503760338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44028431.0, "reward": 0.6374227404594421, "reward_std": 0.14525410532951355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2748437523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.815625011920929, "step": 3735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 766.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 571.3125, "completions/mean_terminated_length": 571.3125, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.7472, "grad_norm": 0.09647620469331741, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44041396.0, "reward": 1.1650009155273438, "reward_std": 0.4382022023200989, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7050000429153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 3736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 477.125, "completions/mean_terminated_length": 440.66668701171875, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.7474, "grad_norm": 0.11085881292819977, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44052838.0, "reward": 1.2673945426940918, "reward_std": 0.45344066619873047, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8472874760627747, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 3737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 803.125, "completions/mean_terminated_length": 771.5714721679688, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.7476, "grad_norm": 0.11029636859893799, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44069816.0, "reward": 1.0406067371368408, "reward_std": 0.5109193325042725, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7687125205993652, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5787500143051147, "step": 3738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 325.75, "completions/mean_terminated_length": 325.75, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.7478, "grad_norm": 0.00017475913045927882, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44078884.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 807.875, "completions/mean_terminated_length": 758.0, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.748, "grad_norm": 0.21725232899188995, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44096994.0, "reward": 0.746719241142273, "reward_std": 0.3997151553630829, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6809375286102295, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.18125000596046448, "step": 3740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 420.4375, "completions/mean_terminated_length": 420.4375, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.7482, "grad_norm": 0.2663375735282898, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44107753.0, "reward": 1.105366826057434, "reward_std": 0.47488635778427124, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6482312083244324, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9193750023841858, "step": 3741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 449.8125, "completions/mean_terminated_length": 449.8125, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.7484, "grad_norm": 0.07627511024475098, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44119430.0, "reward": 1.3764851093292236, "reward_std": 0.279386043548584, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9404684901237488, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.846750020980835, "step": 3742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 393.75, "completions/mean_terminated_length": 393.75, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.7486, "grad_norm": 0.2530295252799988, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44129602.0, "reward": 0.790503978729248, "reward_std": 0.4137791395187378, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33100625872612, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9043750166893005, "step": 3743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 376.125, "completions/mean_terminated_length": 376.125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.7488, "grad_norm": 0.1703699231147766, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44139540.0, "reward": 0.7235947847366333, "reward_std": 0.38415613770484924, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25968748331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 392.125, "completions/mean_terminated_length": 392.125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.749, "grad_norm": 0.37444329261779785, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44149542.0, "reward": 1.1360666751861572, "reward_std": 0.4842212200164795, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6471312642097473, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 3745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 448.0, "completions/mean_terminated_length": 448.0, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.7492, "grad_norm": 0.15323367714881897, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44160694.0, "reward": 1.0230445861816406, "reward_std": 0.4122735559940338, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6710875034332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7587499618530273, "step": 3746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 295.625, "completions/mean_terminated_length": 295.625, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.7494, "grad_norm": 0.2735994756221771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44169328.0, "reward": 0.7748072147369385, "reward_std": 0.4321444630622864, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29961249232292175, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 3747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 844.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 396.375, "completions/mean_terminated_length": 396.375, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.7496, "grad_norm": 0.20506851375102997, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44179558.0, "reward": 1.1173416376113892, "reward_std": 0.42516958713531494, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.734681248664856, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7743750214576721, "step": 3748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 766.4375, "completions/mean_terminated_length": 729.6428833007812, "completions/min_length": 560.0, "completions/min_terminated_length": 560.0, "epoch": 0.7498, "grad_norm": 0.06374054402112961, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44196621.0, "reward": 0.5080476403236389, "reward_std": 0.2272862195968628, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14109374582767487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.778124988079071, "step": 3749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 251.5, "completions/mean_terminated_length": 251.5, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.75, "grad_norm": 0.1303190290927887, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44204229.0, "reward": 1.4375510215759277, "reward_std": 0.2473607212305069, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9376000165939331, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 3750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 227.0, "completions/max_terminated_length": 227.0, "completions/mean_length": 154.5625, "completions/mean_terminated_length": 154.5625, "completions/min_length": 105.0, "completions/min_terminated_length": 105.0, "epoch": 0.7502, "grad_norm": 0.0007948554703034461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44210462.0, "reward": 1.499407410621643, "reward_std": 0.0006015541730448604, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9988124966621399, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 3751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 438.4375, "completions/mean_terminated_length": 438.4375, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.7504, "grad_norm": 0.18966127932071686, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44221205.0, "reward": 0.9078792333602905, "reward_std": 0.47345954179763794, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44075626134872437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 665.0, "completions/mean_terminated_length": 641.0667114257812, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.7506, "grad_norm": 0.0908873900771141, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44235557.0, "reward": 0.571403980255127, "reward_std": 0.28247976303100586, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1428062617778778, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 3753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 586.25, "completions/mean_terminated_length": 557.0667114257812, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.7508, "grad_norm": 0.17106065154075623, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44248809.0, "reward": 0.9244537353515625, "reward_std": 0.47465866804122925, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6614062786102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.581250011920929, "step": 3754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 321.5, "completions/mean_terminated_length": 321.5, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.751, "grad_norm": 0.14879773557186127, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44257393.0, "reward": 0.6002727746963501, "reward_std": 0.24000661075115204, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13804374635219574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 324.0, "completions/max_terminated_length": 324.0, "completions/mean_length": 255.25, "completions/mean_terminated_length": 255.25, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.7512, "grad_norm": 0.0003566559753380716, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44265061.0, "reward": 1.499035358428955, "reward_std": 0.000510936020873487, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980687499046326, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 3756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 948.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 585.0, "completions/mean_terminated_length": 585.0, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.7514, "grad_norm": 0.06353051215410233, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44278661.0, "reward": 0.6596102714538574, "reward_std": 0.25777143239974976, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2567187547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8718749284744263, "step": 3757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 877.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 612.25, "completions/mean_terminated_length": 612.25, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.7516, "grad_norm": 0.11932633817195892, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44292841.0, "reward": 1.132794737815857, "reward_std": 0.4455452561378479, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6405875086784363, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8662500381469727, "step": 3758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 800.875, "completions/mean_terminated_length": 749.3846435546875, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.7518, "grad_norm": 0.0919460877776146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44310743.0, "reward": 0.6125755310058594, "reward_std": 0.3571026027202606, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4126499891281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4962500035762787, "step": 3759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 287.0, "completions/max_terminated_length": 287.0, "completions/mean_length": 206.625, "completions/mean_terminated_length": 206.625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.752, "grad_norm": 0.1429804563522339, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44317809.0, "reward": 0.6621384620666504, "reward_std": 0.32698649168014526, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19927498698234558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 332.125, "completions/mean_terminated_length": 332.125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.7522, "grad_norm": 0.19963492453098297, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44327219.0, "reward": 0.9566197991371155, "reward_std": 0.494331419467926, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4757375121116638, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 3761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 512.0, "completions/mean_terminated_length": 512.0, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.7524, "grad_norm": 0.19578103721141815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44339955.0, "reward": 0.8626446723937988, "reward_std": 0.44807764887809753, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4127874970436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "step": 3762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 375.0, "completions/mean_terminated_length": 375.0, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.7526, "grad_norm": 0.18091724812984467, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44350339.0, "reward": 0.7603133320808411, "reward_std": 0.27478283643722534, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4581250250339508, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 3763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 765.0625, "completions/mean_terminated_length": 728.0714721679688, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.7528, "grad_norm": 0.17925101518630981, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44366612.0, "reward": 0.8463163375854492, "reward_std": 0.5088065266609192, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5051312446594238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7143750190734863, "step": 3764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 614.1875, "completions/mean_terminated_length": 586.86669921875, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.753, "grad_norm": 0.16516728699207306, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44380103.0, "reward": 0.8950791358947754, "reward_std": 0.5159844756126404, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4776562452316284, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 3765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 328.0625, "completions/mean_terminated_length": 328.0625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.7532, "grad_norm": 0.24045029282569885, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44389608.0, "reward": 1.2317196130752563, "reward_std": 0.4205136001110077, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7759374976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 3766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 421.0625, "completions/mean_terminated_length": 421.0625, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.7534, "grad_norm": 0.004597336985170841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44400537.0, "reward": 0.5498322248458862, "reward_std": 0.014043881557881832, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09966249763965607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 3767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 593.9375, "completions/mean_terminated_length": 593.9375, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.7536, "grad_norm": 0.09345975518226624, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44414168.0, "reward": 1.408729076385498, "reward_std": 0.26097220182418823, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9424562454223633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 3768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 537.8125, "completions/mean_terminated_length": 505.4000244140625, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.7538, "grad_norm": 0.03959647938609123, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44427509.0, "reward": 0.5287508964538574, "reward_std": 0.14358381927013397, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11999999731779099, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.875, "step": 3769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 279.5625, "completions/mean_terminated_length": 279.5625, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.754, "grad_norm": 0.10837731510400772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44435614.0, "reward": 0.7697663307189941, "reward_std": 0.2162407636642456, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5395312309265137, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5406249761581421, "step": 3770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 300.625, "completions/mean_terminated_length": 300.625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.7542, "grad_norm": 0.09610477834939957, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44444712.0, "reward": 1.4370291233062744, "reward_std": 0.24722103774547577, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9365562200546265, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 622.625, "completions/mean_terminated_length": 595.86669921875, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.7544, "grad_norm": 0.15422563254833221, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44458386.0, "reward": 1.0538289546966553, "reward_std": 0.5063629746437073, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6701562404632568, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.765625, "step": 3772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 799.375, "completions/mean_terminated_length": 767.2857666015625, "completions/min_length": 585.0, "completions/min_terminated_length": 585.0, "epoch": 0.7546, "grad_norm": 0.05908369645476341, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44475896.0, "reward": 0.4849226474761963, "reward_std": 0.25709307193756104, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15734374523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 3773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 685.25, "completions/mean_terminated_length": 662.6666870117188, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.7548, "grad_norm": 0.08535565435886383, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44491084.0, "reward": 0.773737907409668, "reward_std": 0.2885546088218689, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6099750399589539, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4424999952316284, "step": 3774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 856.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 681.4375, "completions/mean_terminated_length": 681.4375, "completions/min_length": 511.0, "completions/min_terminated_length": 511.0, "epoch": 0.755, "grad_norm": 0.10017897188663483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44507491.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 417.625, "completions/mean_terminated_length": 417.625, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.7552, "grad_norm": 0.0305799450725317, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44518045.0, "reward": 1.467482089996338, "reward_std": 0.12599605321884155, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9974625110626221, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 3776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 677.75, "completions/mean_terminated_length": 677.75, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.7554, "grad_norm": 0.14564967155456543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44533593.0, "reward": 0.8076696395874023, "reward_std": 0.4226558804512024, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3653374910354614, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 3777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 926.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 677.625, "completions/mean_terminated_length": 677.625, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.7556, "grad_norm": 0.17386935651302338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44548323.0, "reward": 0.9642196893692017, "reward_std": 0.4871813952922821, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4909375011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 717.3125, "completions/mean_terminated_length": 673.5, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.7558, "grad_norm": 0.21212394535541534, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44563752.0, "reward": 0.9784382581710815, "reward_std": 0.5760053992271423, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5818750262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.800000011920929, "step": 3779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 427.6875, "completions/mean_terminated_length": 427.6875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.756, "grad_norm": 0.2073001265525818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44574675.0, "reward": 1.377866506576538, "reward_std": 0.3313027322292328, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8807312846183777, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 3780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 462.625, "completions/mean_terminated_length": 462.625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.7562, "grad_norm": 0.13076038658618927, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44586893.0, "reward": 0.7226415872573853, "reward_std": 0.38523033261299133, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2577812373638153, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 814.625, "completions/mean_terminated_length": 784.7142944335938, "completions/min_length": 620.0, "completions/min_terminated_length": 620.0, "epoch": 0.7564, "grad_norm": 0.12044793367385864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44605127.0, "reward": 0.7376567125320435, "reward_std": 0.34146952629089355, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6003124713897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3812499940395355, "step": 3782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 465.375, "completions/mean_terminated_length": 465.375, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.7566, "grad_norm": 0.15006123483181, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44616189.0, "reward": 0.7267947196960449, "reward_std": 0.38358861207962036, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26608750224113464, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 408.375, "completions/mean_terminated_length": 408.375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.7568, "grad_norm": 0.14918793737888336, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44626947.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 921.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 545.0625, "completions/mean_terminated_length": 545.0625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.757, "grad_norm": 0.19950933754444122, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44639396.0, "reward": 1.0288288593292236, "reward_std": 0.43024566769599915, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6201562285423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.809374988079071, "step": 3785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 901.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 748.375, "completions/mean_terminated_length": 748.375, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.7572, "grad_norm": 0.1317971795797348, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44655498.0, "reward": 0.5892195701599121, "reward_std": 0.19695264101028442, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2409374862909317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 3786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 453.25, "completions/mean_terminated_length": 453.25, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.7574, "grad_norm": 0.1565500646829605, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44666622.0, "reward": 0.7302196025848389, "reward_std": 0.37144607305526733, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27293750643730164, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9199999570846558, "step": 3787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 670.5, "completions/mean_terminated_length": 670.5, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.7576, "grad_norm": 0.06574622541666031, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44681558.0, "reward": 0.8121100664138794, "reward_std": 0.2330859899520874, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.561718761920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.659375011920929, "step": 3788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 819.125, "completions/mean_terminated_length": 750.8333740234375, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.7578, "grad_norm": 0.10972315818071365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44699720.0, "reward": 0.5731254816055298, "reward_std": 0.3731801509857178, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.39625000953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.4624999761581421, "step": 3789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 904.375, "completions/mean_terminated_length": 750.5714721679688, "completions/min_length": 579.0, "completions/min_terminated_length": 579.0, "epoch": 0.758, "grad_norm": 0.2813836336135864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44719150.0, "reward": 0.4074222445487976, "reward_std": 0.5480653047561646, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2523437738418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.328125, "step": 3790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 247.75, "completions/mean_terminated_length": 247.75, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.7582, "grad_norm": 0.020615125074982643, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44726666.0, "reward": 0.5524165630340576, "reward_std": 0.02947871945798397, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10483124852180481, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 3791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 616.9375, "completions/mean_terminated_length": 616.9375, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.7584, "grad_norm": 0.14558757841587067, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44741225.0, "reward": 1.0059382915496826, "reward_std": 0.45645588636398315, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5743750333786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 3792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 827.1875, "completions/mean_terminated_length": 799.0714721679688, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.7586, "grad_norm": 0.21173536777496338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44758780.0, "reward": 0.7857006788253784, "reward_std": 0.4579770863056183, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5089000463485718, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6025000214576721, "step": 3793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 822.4375, "completions/mean_terminated_length": 775.923095703125, "completions/min_length": 605.0, "completions/min_terminated_length": 605.0, "epoch": 0.7588, "grad_norm": 0.2191348671913147, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44776835.0, "reward": 0.6621880531311035, "reward_std": 0.37916404008865356, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5118750333786011, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.3812499940395355, "step": 3794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 741.875, "completions/mean_terminated_length": 723.0667114257812, "completions/min_length": 621.0, "completions/min_terminated_length": 621.0, "epoch": 0.759, "grad_norm": 0.18258118629455566, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44792609.0, "reward": 0.989672839641571, "reward_std": 0.5426595211029053, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5418437719345093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981249928474426, "step": 3795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 445.0, "completions/max_terminated_length": 445.0, "completions/mean_length": 314.5, "completions/mean_terminated_length": 314.5, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.7592, "grad_norm": 0.4428834617137909, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44802025.0, "reward": 1.017735481262207, "reward_std": 0.4971912205219269, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5354687571525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 3796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 952.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 419.0, "completions/mean_terminated_length": 419.0, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.7594, "grad_norm": 0.19245587289333344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44812409.0, "reward": 0.8582820296287537, "reward_std": 0.40925332903862, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.46656250953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8249999284744263, "step": 3797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 491.4375, "completions/mean_terminated_length": 455.933349609375, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.7596, "grad_norm": 0.13508544862270355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44824480.0, "reward": 0.7170319557189941, "reward_std": 0.3671148419380188, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.37156251072883606, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.75, "step": 3798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 602.375, "completions/mean_terminated_length": 542.1428833007812, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.7598, "grad_norm": 0.18508459627628326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44838326.0, "reward": 0.7026569843292236, "reward_std": 0.3668506145477295, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.46781250834465027, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5874999761581421, "step": 3799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 789.125, "completions/mean_terminated_length": 734.923095703125, "completions/min_length": 553.0, "completions/min_terminated_length": 553.0, "epoch": 0.76, "grad_norm": 0.11526205390691757, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44855304.0, "reward": 0.542697548866272, "reward_std": 0.3131338953971863, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2728937268257141, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6181249618530273, "step": 3800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 278.0, "completions/mean_terminated_length": 278.0, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.7602, "grad_norm": 0.4049726724624634, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44863192.0, "reward": 1.264347791671753, "reward_std": 0.41993260383605957, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7786937952041626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 617.875, "completions/mean_terminated_length": 590.800048828125, "completions/min_length": 454.0, "completions/min_terminated_length": 454.0, "epoch": 0.7604, "grad_norm": 0.05909629538655281, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44877174.0, "reward": 0.5361696481704712, "reward_std": 0.15492777526378632, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13483749330043793, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8650000095367432, "step": 3802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 490.75, "completions/mean_terminated_length": 490.75, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.7606, "grad_norm": 0.2860282361507416, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44888978.0, "reward": 0.8163288831710815, "reward_std": 0.37979164719581604, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4451562464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.784375011920929, "step": 3803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 546.0, "completions/mean_terminated_length": 546.0, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.7608, "grad_norm": 0.06010820344090462, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44902498.0, "reward": 0.5832823514938354, "reward_std": 0.12432941794395447, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1665624976158142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 3804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 345.3125, "completions/mean_terminated_length": 345.3125, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.761, "grad_norm": 0.22434379160404205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44912135.0, "reward": 1.321341633796692, "reward_std": 0.3833126127719879, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8301812410354614, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 3805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 829.75, "completions/mean_terminated_length": 784.923095703125, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.7612, "grad_norm": 0.1865173727273941, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44929827.0, "reward": 0.653984785079956, "reward_std": 0.3676551282405853, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49546873569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4281249940395355, "step": 3806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 882.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 653.5, "completions/mean_terminated_length": 653.5, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.7614, "grad_norm": 0.16099636256694794, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44944603.0, "reward": 1.0807009935379028, "reward_std": 0.4896578788757324, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5989000201225281, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 377.8125, "completions/mean_terminated_length": 377.8125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.7616, "grad_norm": 0.4134197235107422, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44954408.0, "reward": 1.0423321723937988, "reward_std": 0.47814348340034485, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5846624970436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 3808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 301.0, "completions/max_terminated_length": 301.0, "completions/mean_length": 261.8125, "completions/mean_terminated_length": 261.8125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.7618, "grad_norm": 0.22632859647274017, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44962853.0, "reward": 1.3776509761810303, "reward_std": 0.33121833205223083, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8803000450134277, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 3809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 378.1875, "completions/mean_terminated_length": 378.1875, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.762, "grad_norm": 0.2365289330482483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44972840.0, "reward": 1.0242195129394531, "reward_std": 0.40563368797302246, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6109374761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 3810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 380.5625, "completions/mean_terminated_length": 380.5625, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.7622, "grad_norm": 0.28065577149391174, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44982273.0, "reward": 0.6656915545463562, "reward_std": 0.3252728283405304, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20638124644756317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 266.0, "completions/max_terminated_length": 266.0, "completions/mean_length": 204.8125, "completions/mean_terminated_length": 204.8125, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.7624, "grad_norm": 0.013771923258900642, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44989438.0, "reward": 0.5432791709899902, "reward_std": 0.01506174635142088, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08655624836683273, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249380111694, "step": 3812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 365.75, "completions/mean_terminated_length": 365.75, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.7626, "grad_norm": 0.132585808634758, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44998938.0, "reward": 0.6525789499282837, "reward_std": 0.18087467551231384, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.30515626072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7718750238418579, "step": 3813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 447.125, "completions/mean_terminated_length": 447.125, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.7628, "grad_norm": 0.14509737491607666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45010076.0, "reward": 0.6509009599685669, "reward_std": 0.29809239506721497, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1767999827861786, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9212499856948853, "step": 3814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 606.5625, "completions/mean_terminated_length": 578.7333374023438, "completions/min_length": 435.0, "completions/min_terminated_length": 435.0, "epoch": 0.763, "grad_norm": 0.11450973153114319, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45023637.0, "reward": 0.6531257629394531, "reward_std": 0.25922393798828125, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.36875003576278687, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 3815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 338.5, "completions/mean_terminated_length": 292.8000183105469, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.7632, "grad_norm": 0.3733304738998413, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45032781.0, "reward": 1.166207194328308, "reward_std": 0.5255309343338013, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7074124813079834, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9075000286102295, "step": 3816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 354.3125, "completions/mean_terminated_length": 354.3125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.7634, "grad_norm": 0.2165028303861618, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45042370.0, "reward": 0.6849040985107422, "reward_std": 0.3230895698070526, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24480624496936798, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9318749904632568, "step": 3817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 206.0, "completions/max_terminated_length": 206.0, "completions/mean_length": 162.3125, "completions/mean_terminated_length": 162.3125, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "epoch": 0.7636, "grad_norm": 0.8530493974685669, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45048663.0, "reward": 0.7332541942596436, "reward_std": 0.3910166025161743, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2790062427520752, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9443750381469727, "step": 3818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 399.5, "completions/mean_terminated_length": 399.5, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.7638, "grad_norm": 0.3140089213848114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45058671.0, "reward": 0.8886637687683105, "reward_std": 0.48856067657470703, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40232497453689575, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9699999690055847, "step": 3819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 565.625, "completions/mean_terminated_length": 565.625, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.764, "grad_norm": 0.1367240846157074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45071609.0, "reward": 1.00632905960083, "reward_std": 0.46213096380233765, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5751562714576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 3820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 859.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 446.3125, "completions/mean_terminated_length": 446.3125, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.7642, "grad_norm": 0.15462900698184967, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45082542.0, "reward": 0.6038916110992432, "reward_std": 0.2389136105775833, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14528124034404755, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 324.5, "completions/mean_terminated_length": 324.5, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.7644, "grad_norm": 0.2088056355714798, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45091382.0, "reward": 1.4379295110702515, "reward_std": 0.24746128916740417, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9383562803268433, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9881250262260437, "step": 3822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 282.75, "completions/mean_terminated_length": 282.75, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.7646, "grad_norm": 0.13144086301326752, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45099714.0, "reward": 1.459869623184204, "reward_std": 0.12777499854564667, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9822374582290649, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8787499666213989, "step": 3823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 679.0625, "completions/mean_terminated_length": 629.7857666015625, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.7648, "grad_norm": 0.06363977491855621, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45114323.0, "reward": 0.47711026668548584, "reward_std": 0.18663929402828217, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 3824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 471.875, "completions/mean_terminated_length": 471.875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.765, "grad_norm": 0.18693895637989044, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45126049.0, "reward": 0.7405480146408081, "reward_std": 0.3882680833339691, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29359376430511475, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 3825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 251.375, "completions/mean_terminated_length": 251.375, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.7652, "grad_norm": 0.3888898193836212, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45134231.0, "reward": 1.3207042217254639, "reward_std": 0.382996141910553, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8289062976837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 460.1875, "completions/mean_terminated_length": 460.1875, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.7654, "grad_norm": 0.2694655954837799, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45145674.0, "reward": 0.8345164060592651, "reward_std": 0.383344441652298, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4815312325954437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7356249690055847, "step": 3827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 550.0, "completions/max_terminated_length": 550.0, "completions/mean_length": 439.125, "completions/mean_terminated_length": 439.125, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.7656, "grad_norm": 0.169391468167305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45156476.0, "reward": 1.2643635272979736, "reward_std": 0.38645994663238525, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7787250280380249, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 3828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 511.3125, "completions/mean_terminated_length": 438.0714416503906, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.7658, "grad_norm": 0.30909982323646545, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45168977.0, "reward": 0.8906134366989136, "reward_std": 0.5798611640930176, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4687249958515167, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8362500071525574, "step": 3829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 245.0, "completions/max_terminated_length": 245.0, "completions/mean_length": 188.5, "completions/mean_terminated_length": 188.5, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.766, "grad_norm": 0.026805315166711807, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45175897.0, "reward": 0.5407478213310242, "reward_std": 0.022070741280913353, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08149375021457672, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 865.625, "completions/mean_terminated_length": 770.6000366210938, "completions/min_length": 555.0, "completions/min_terminated_length": 555.0, "epoch": 0.7662, "grad_norm": 0.2937007546424866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45193763.0, "reward": 0.6398442983627319, "reward_std": 0.6422095894813538, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3421874940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.59375, "step": 3831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 533.5625, "completions/mean_terminated_length": 533.5625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.7664, "grad_norm": 0.1435372680425644, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45206796.0, "reward": 0.6810165643692017, "reward_std": 0.3220902383327484, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2370312362909317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 457.6875, "completions/mean_terminated_length": 457.6875, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.7666, "grad_norm": 0.27391430735588074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45218903.0, "reward": 0.8905918002128601, "reward_std": 0.48773568868637085, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4061812162399292, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9755625128746033, "step": 3833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 634.1875, "completions/mean_terminated_length": 578.5, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.7668, "grad_norm": 0.17640598118305206, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45233114.0, "reward": 0.805547297000885, "reward_std": 0.3302585482597351, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7360937595367432, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.265625, "step": 3834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 607.0, "completions/mean_terminated_length": 579.2000122070312, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.767, "grad_norm": 0.14088809490203857, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45246922.0, "reward": 0.8188285231590271, "reward_std": 0.275430828332901, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7001562714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.37187498807907104, "step": 3835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 277.0, "completions/max_terminated_length": 277.0, "completions/mean_length": 221.875, "completions/mean_terminated_length": 221.875, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.7672, "grad_norm": 0.3007798492908478, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45254024.0, "reward": 1.4376230239868164, "reward_std": 0.24737988412380219, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9377436637878418, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9755625128746033, "step": 3836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 392.0625, "completions/mean_terminated_length": 392.0625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.7674, "grad_norm": 0.21042108535766602, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45264137.0, "reward": 1.0832009315490723, "reward_std": 0.48679855465888977, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6039000153541565, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 812.875, "completions/mean_terminated_length": 764.1538696289062, "completions/min_length": 568.0, "completions/min_terminated_length": 568.0, "epoch": 0.7676, "grad_norm": 0.16765353083610535, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45281479.0, "reward": 0.5157039165496826, "reward_std": 0.3485039174556732, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1564062535762787, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7593749761581421, "step": 3838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 509.375, "completions/mean_terminated_length": 475.0666809082031, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.7678, "grad_norm": 0.33649942278862, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45293565.0, "reward": 0.9923445582389832, "reward_std": 0.5400531888008118, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5471875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 3839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 369.4375, "completions/mean_terminated_length": 369.4375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.768, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45303620.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 603.0, "completions/mean_terminated_length": 574.933349609375, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.7682, "grad_norm": 0.10430587828159332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45317684.0, "reward": 0.7437506318092346, "reward_std": 0.269948273897171, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.550000011920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.49375003576278687, "step": 3841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 707.125, "completions/mean_terminated_length": 707.125, "completions/min_length": 561.0, "completions/min_terminated_length": 561.0, "epoch": 0.7684, "grad_norm": 0.09310755133628845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45333126.0, "reward": 0.6757040619850159, "reward_std": 0.322939932346344, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2264062464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 3842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 631.125, "completions/mean_terminated_length": 631.125, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.7686, "grad_norm": 0.2113150954246521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45347192.0, "reward": 1.0766416788101196, "reward_std": 0.4937414824962616, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5907812118530273, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 630.0625, "completions/mean_terminated_length": 603.800048828125, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.7688, "grad_norm": 0.22030098736286163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45361145.0, "reward": 1.0747509002685547, "reward_std": 0.5219303369522095, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6495000123977661, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8637499809265137, "step": 3844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 392.1875, "completions/mean_terminated_length": 392.1875, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.769, "grad_norm": 0.24196204543113708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45371116.0, "reward": 1.026641607284546, "reward_std": 0.48772040009498596, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5532812476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 760.9375, "completions/mean_terminated_length": 723.357177734375, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.7692, "grad_norm": 0.055949460715055466, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45387387.0, "reward": 0.5092726349830627, "reward_std": 0.22645805776119232, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14354375004768372, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7818750143051147, "step": 3846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 404.5, "completions/mean_terminated_length": 363.20001220703125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.7694, "grad_norm": 0.4403279721736908, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45397603.0, "reward": 0.9511727094650269, "reward_std": 0.5047948956489563, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.52734375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.784375011920929, "step": 3847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 434.75, "completions/mean_terminated_length": 434.75, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.7696, "grad_norm": 0.3363742530345917, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45408351.0, "reward": 1.1559321880340576, "reward_std": 0.46005508303642273, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6868624687194824, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362500309944153, "step": 3848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 401.3125, "completions/mean_terminated_length": 401.3125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.7698, "grad_norm": 0.06818623840808868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45419044.0, "reward": 1.466719627380371, "reward_std": 0.12448550015687943, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9959374666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 3849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 299.0625, "completions/mean_terminated_length": 299.0625, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.77, "grad_norm": 0.00026057768263854086, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45427957.0, "reward": 1.4990540742874146, "reward_std": 0.0005425068666227162, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981062412261963, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 3850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 515.5, "completions/mean_terminated_length": 515.5, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.7702, "grad_norm": 0.15946803987026215, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45440045.0, "reward": 0.7612508535385132, "reward_std": 0.19346904754638672, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5225000381469727, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.606249988079071, "step": 3851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 267.6875, "completions/mean_terminated_length": 267.6875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.7704, "grad_norm": 0.2927691340446472, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45448488.0, "reward": 0.5978041291236877, "reward_std": 0.24095547199249268, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13310624659061432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606250524520874, "step": 3852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 784.0, "completions/mean_terminated_length": 749.7142944335938, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.7706, "grad_norm": 0.2720293700695038, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45465192.0, "reward": 0.7682035565376282, "reward_std": 0.34500035643577576, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6614062786102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 3853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 369.8125, "completions/mean_terminated_length": 369.8125, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.7708, "grad_norm": 0.2448444962501526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45475397.0, "reward": 1.3785479068756104, "reward_std": 0.3296127915382385, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8820937275886536, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 346.75, "completions/mean_terminated_length": 346.75, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.771, "grad_norm": 0.1651308387517929, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45484593.0, "reward": 1.0296099185943604, "reward_std": 0.31246134638786316, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7467187643051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.528124988079071, "step": 3855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 338.0, "completions/max_terminated_length": 338.0, "completions/mean_length": 252.3125, "completions/mean_terminated_length": 252.3125, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.7712, "grad_norm": 0.3821021616458893, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45492278.0, "reward": 1.3800666332244873, "reward_std": 0.3245124816894531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 572.5625, "completions/mean_terminated_length": 572.5625, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.7714, "grad_norm": 0.11370758712291718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45505855.0, "reward": 0.9058600068092346, "reward_std": 0.34149807691574097, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.624218761920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6593750715255737, "step": 3857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 301.0, "completions/max_terminated_length": 301.0, "completions/mean_length": 260.0, "completions/mean_terminated_length": 260.0, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.7716, "grad_norm": 0.0006127910455688834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45513839.0, "reward": 1.4991728067398071, "reward_std": 0.0005645550554618239, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983437657356262, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 744.0, "completions/mean_terminated_length": 704.0000610351562, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.7718, "grad_norm": 0.19146129488945007, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45529279.0, "reward": 0.6071600914001465, "reward_std": 0.358140230178833, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.27681875228881836, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7156250476837158, "step": 3859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 509.25, "completions/mean_terminated_length": 509.25, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.772, "grad_norm": 0.23634211719036102, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45540963.0, "reward": 0.8029696941375732, "reward_std": 0.4166632294654846, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35593751072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 3860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 552.5625, "completions/mean_terminated_length": 521.1333618164062, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.7722, "grad_norm": 0.3856232762336731, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45553548.0, "reward": 0.8935163617134094, "reward_std": 0.4729250371456146, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5370312333106995, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.765625, "step": 3861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 466.625, "completions/mean_terminated_length": 466.625, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.7724, "grad_norm": 0.1774539202451706, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45565094.0, "reward": 0.7741414904594421, "reward_std": 0.3710496723651886, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3607812523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 3862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 482.1875, "completions/mean_terminated_length": 482.1875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.7726, "grad_norm": 0.26970377564430237, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45576249.0, "reward": 1.088750958442688, "reward_std": 0.48062458634376526, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6150000095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 3863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 496.4375, "completions/mean_terminated_length": 496.4375, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.7728, "grad_norm": 0.42185625433921814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45587952.0, "reward": 0.9702321887016296, "reward_std": 0.48178255558013916, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5029624700546265, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "step": 3864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 477.625, "completions/mean_terminated_length": 477.625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.773, "grad_norm": 0.1430649608373642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45599226.0, "reward": 0.8217192888259888, "reward_std": 0.24184849858283997, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5809375047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5812499523162842, "step": 3865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 379.9375, "completions/mean_terminated_length": 379.9375, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.7732, "grad_norm": 0.23997730016708374, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45609609.0, "reward": 1.1731133460998535, "reward_std": 0.4460124969482422, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7212250232696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 554.75, "completions/mean_terminated_length": 487.71429443359375, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.7734, "grad_norm": 0.41465216875076294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45622341.0, "reward": 0.5770319700241089, "reward_std": 0.3258732557296753, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21656250953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.762499988079071, "step": 3867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 690.8125, "completions/mean_terminated_length": 643.2142944335938, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.7736, "grad_norm": 0.18893182277679443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45637170.0, "reward": 0.7406131029129028, "reward_std": 0.38548579812049866, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5437250137329102, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.48000001907348633, "step": 3868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 829.1875, "completions/mean_terminated_length": 740.6364135742188, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.7738, "grad_norm": 0.12085496634244919, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45654965.0, "reward": 0.4031912684440613, "reward_std": 0.3023384213447571, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1188812404870987, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6018750071525574, "step": 3869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 537.0, "completions/max_terminated_length": 537.0, "completions/mean_length": 393.8125, "completions/mean_terminated_length": 393.8125, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.774, "grad_norm": 0.20094603300094604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45665426.0, "reward": 1.2014071941375732, "reward_std": 0.45445317029953003, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7153124809265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 3870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 307.5625, "completions/mean_terminated_length": 307.5625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.7742, "grad_norm": 0.0933433547616005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45675579.0, "reward": 1.439507246017456, "reward_std": 0.2375352531671524, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 3871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 790.5625, "completions/mean_terminated_length": 775.0000610351562, "completions/min_length": 561.0, "completions/min_terminated_length": 561.0, "epoch": 0.7744, "grad_norm": 0.22768163681030273, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45692532.0, "reward": 1.1145195960998535, "reward_std": 0.4319396913051605, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.8540375232696533, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5425000190734863, "step": 3872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 871.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 484.875, "completions/mean_terminated_length": 484.875, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.7746, "grad_norm": 0.2363254874944687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45704194.0, "reward": 1.246082067489624, "reward_std": 0.46166491508483887, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.7421624660491943, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 3873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 528.1875, "completions/mean_terminated_length": 528.1875, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.7748, "grad_norm": 0.21411731839179993, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45717797.0, "reward": 1.3485946655273438, "reward_std": 0.3359578549861908, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8846874833106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 3874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 464.75, "completions/mean_terminated_length": 464.75, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.775, "grad_norm": 0.1909378170967102, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45728945.0, "reward": 0.726875901222229, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26625001430511475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 531.75, "completions/mean_terminated_length": 531.75, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.7752, "grad_norm": 0.13310520350933075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45741005.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 433.9375, "completions/mean_terminated_length": 433.9375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.7754, "grad_norm": 0.17207829654216766, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45751484.0, "reward": 0.6102884411811829, "reward_std": 0.2377483993768692, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1580749899148941, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 3877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 949.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 619.4375, "completions/mean_terminated_length": 619.4375, "completions/min_length": 481.0, "completions/min_terminated_length": 481.0, "epoch": 0.7756, "grad_norm": 0.16333593428134918, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45765219.0, "reward": 1.1425416469573975, "reward_std": 0.4756823778152466, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600812673568726, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 637.0, "completions/max_terminated_length": 637.0, "completions/mean_length": 517.25, "completions/mean_terminated_length": 517.25, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.7758, "grad_norm": 0.3353157043457031, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45777303.0, "reward": 0.7472665309906006, "reward_std": 0.38152357935905457, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3070312440395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 3879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 572.75, "completions/mean_terminated_length": 542.6666870117188, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.776, "grad_norm": 0.170829176902771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45790787.0, "reward": 0.90140700340271, "reward_std": 0.4596075117588043, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5528124570846558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7499999403953552, "step": 3880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 482.25, "completions/mean_terminated_length": 482.25, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.7762, "grad_norm": 0.17956046760082245, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45803175.0, "reward": 0.9702353477478027, "reward_std": 0.48178631067276, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5029687881469727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 642.5, "completions/mean_terminated_length": 617.0667114257812, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.7764, "grad_norm": 0.29615989327430725, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45817839.0, "reward": 0.690153956413269, "reward_std": 0.4234946668148041, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2553062438964844, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 3882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 505.8125, "completions/mean_terminated_length": 471.2666931152344, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.7766, "grad_norm": 0.13556501269340515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45829740.0, "reward": 1.2438163757324219, "reward_std": 0.447622150182724, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8626312613487244, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7581249475479126, "step": 3883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 301.6875, "completions/mean_terminated_length": 301.6875, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.7768, "grad_norm": 0.32549381256103516, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45837975.0, "reward": 1.0213258266448975, "reward_std": 0.49317145347595215, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5426499843597412, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 473.3125, "completions/mean_terminated_length": 436.60003662109375, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.777, "grad_norm": 0.16446273028850555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45849356.0, "reward": 0.7560665011405945, "reward_std": 0.4642445743083954, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3246312737464905, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 3885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 438.8125, "completions/mean_terminated_length": 355.21429443359375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.7772, "grad_norm": 0.5570343136787415, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45860249.0, "reward": 1.0586724281311035, "reward_std": 0.5252925157546997, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7423437237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 3886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 720.625, "completions/mean_terminated_length": 650.6154174804688, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.7774, "grad_norm": 0.16677957773208618, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45875475.0, "reward": 0.8149038553237915, "reward_std": 0.5923504829406738, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4423062801361084, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7568750381469727, "step": 3887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 366.5, "completions/mean_terminated_length": 366.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.7776, "grad_norm": 0.6416021585464478, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45885035.0, "reward": 1.1057790517807007, "reward_std": 0.5393296480178833, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6490562558174133, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 399.8125, "completions/mean_terminated_length": 399.8125, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.7778, "grad_norm": 0.07556465268135071, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45895416.0, "reward": 1.4376978874206543, "reward_std": 0.2447727471590042, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9378937482833862, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 351.125, "completions/mean_terminated_length": 351.125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.778, "grad_norm": 0.34602949023246765, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45905514.0, "reward": 1.0068638324737549, "reward_std": 0.5085405707359314, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5137250423431396, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 3890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 594.125, "completions/mean_terminated_length": 565.4666748046875, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.7782, "grad_norm": 0.09980936348438263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45921628.0, "reward": 0.6935102343559265, "reward_std": 0.4227396547794342, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26201874017715454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 3891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 277.4375, "completions/mean_terminated_length": 277.4375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.7784, "grad_norm": 0.4633718729019165, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45929891.0, "reward": 1.2782697677612305, "reward_std": 0.39920011162757874, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8065375089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "step": 3892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 446.8125, "completions/mean_terminated_length": 446.8125, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.7786, "grad_norm": 0.20681273937225342, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45941232.0, "reward": 1.1708476543426514, "reward_std": 0.4473702311515808, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7166937589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081250429153442, "step": 3893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 426.875, "completions/mean_terminated_length": 426.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.7788, "grad_norm": 0.26774200797080994, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45952270.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 692.8125, "completions/mean_terminated_length": 616.3846435546875, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.779, "grad_norm": 0.1163337230682373, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45967595.0, "reward": 0.8507816791534424, "reward_std": 0.47488224506378174, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7015625238418579, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.26249998807907104, "step": 3895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 668.0625, "completions/mean_terminated_length": 668.0625, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.7792, "grad_norm": 0.21108180284500122, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45982764.0, "reward": 0.6610165238380432, "reward_std": 0.3606753349304199, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25953125953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.859375, "step": 3896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 413.5, "completions/mean_terminated_length": 413.5, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.7794, "grad_norm": 0.172106072306633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45993124.0, "reward": 0.7457603216171265, "reward_std": 0.3910130560398102, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3040187358856201, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 3897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 497.875, "completions/mean_terminated_length": 497.875, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.7796, "grad_norm": 0.16470886766910553, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46006354.0, "reward": 0.7431915402412415, "reward_std": 0.3778519332408905, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2988812327384949, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 3898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 465.75, "completions/mean_terminated_length": 465.75, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.7798, "grad_norm": 0.21711735427379608, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46018206.0, "reward": 1.1833571195602417, "reward_std": 0.4317837953567505, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7417125105857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 3899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 436.625, "completions/mean_terminated_length": 436.625, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.78, "grad_norm": 0.2850421071052551, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46028936.0, "reward": 0.8432009220123291, "reward_std": 0.45656633377075195, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37389999628067017, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 654.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 526.3125, "completions/mean_terminated_length": 526.3125, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.7802, "grad_norm": 0.1168164610862732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46042733.0, "reward": 0.7267976999282837, "reward_std": 0.2719525992870331, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.39109376072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.765625, "step": 3901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 255.6875, "completions/mean_terminated_length": 255.6875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.7804, "grad_norm": 0.0006798555259592831, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46050552.0, "reward": 1.4991071224212646, "reward_std": 0.0005474113277159631, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982125163078308, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 3902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1003.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 559.5625, "completions/mean_terminated_length": 559.5625, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.7806, "grad_norm": 0.21930043399333954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46065089.0, "reward": 0.9314446449279785, "reward_std": 0.46120962500572205, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.48788750171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 3903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 437.8125, "completions/mean_terminated_length": 437.8125, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.7808, "grad_norm": 0.07229848951101303, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46076190.0, "reward": 0.6057009696960449, "reward_std": 0.2383432537317276, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14890000224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 3904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 425.6875, "completions/mean_terminated_length": 425.6875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.781, "grad_norm": 0.3577243983745575, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46087033.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 524.125, "completions/mean_terminated_length": 524.125, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.7812, "grad_norm": 0.07493928074836731, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46099003.0, "reward": 0.8166412115097046, "reward_std": 0.14525800943374634, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6332812309265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.546875, "step": 3906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 369.9375, "completions/mean_terminated_length": 369.9375, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.7814, "grad_norm": 0.4995782971382141, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46108666.0, "reward": 0.8749227523803711, "reward_std": 0.43863484263420105, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4373437464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 3907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 609.0, "completions/mean_terminated_length": 581.3333740234375, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.7816, "grad_norm": 0.10500229150056839, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46122474.0, "reward": 0.6313289403915405, "reward_std": 0.30263999104499817, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.26265624165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.809374988079071, "step": 3908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 687.875, "completions/mean_terminated_length": 610.3077392578125, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.7818, "grad_norm": 0.23167207837104797, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46138104.0, "reward": 0.7198441028594971, "reward_std": 0.37544453144073486, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6271874904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 3909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 321.1875, "completions/mean_terminated_length": 321.1875, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.782, "grad_norm": 0.205204576253891, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46147787.0, "reward": 0.8427348136901855, "reward_std": 0.16535425186157227, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6854687929153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.46562501788139343, "step": 3910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 591.5625, "completions/mean_terminated_length": 591.5625, "completions/min_length": 436.0, "completions/min_terminated_length": 436.0, "epoch": 0.7822, "grad_norm": 0.027203606441617012, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46161124.0, "reward": 0.6102352142333984, "reward_std": 0.0782679095864296, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22046874463558197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 3911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 407.4375, "completions/mean_terminated_length": 366.3333435058594, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.7824, "grad_norm": 0.514019250869751, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46171547.0, "reward": 1.21781325340271, "reward_std": 0.46992871165275574, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8106250166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7624999284744263, "step": 3912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 877.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 546.25, "completions/mean_terminated_length": 546.25, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.7826, "grad_norm": 0.1843976080417633, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46184191.0, "reward": 0.9466415643692017, "reward_std": 0.43608948588371277, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5182812213897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 3913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 770.1875, "completions/mean_terminated_length": 753.2667236328125, "completions/min_length": 535.0, "completions/min_terminated_length": 535.0, "epoch": 0.7828, "grad_norm": 0.08145233988761902, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46200818.0, "reward": 0.7085912227630615, "reward_std": 0.27896496653556824, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47968125343322754, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5493749976158142, "step": 3914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 309.0, "completions/max_terminated_length": 309.0, "completions/mean_length": 240.3125, "completions/mean_terminated_length": 240.3125, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.783, "grad_norm": 0.3757697641849518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46208119.0, "reward": 0.8359915018081665, "reward_std": 0.4619092643260956, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35948124527931213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 3915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 557.6875, "completions/mean_terminated_length": 557.6875, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.7832, "grad_norm": 0.08111798763275146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46221026.0, "reward": 0.7608569860458374, "reward_std": 0.17432136833667755, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5217124819755554, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6399999856948853, "step": 3916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 627.6875, "completions/mean_terminated_length": 627.6875, "completions/min_length": 522.0, "completions/min_terminated_length": 522.0, "epoch": 0.7834, "grad_norm": 0.05816841125488281, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46234781.0, "reward": 1.3456759452819824, "reward_std": 0.3443696200847626, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.878849983215332, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 3917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 485.6875, "completions/mean_terminated_length": 485.6875, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.7836, "grad_norm": 0.2417633980512619, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46246472.0, "reward": 0.8712508678436279, "reward_std": 0.44184410572052, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4300000071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 3918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 896.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 728.375, "completions/mean_terminated_length": 728.375, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.7838, "grad_norm": 0.046338461339473724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46262174.0, "reward": 0.5732821822166443, "reward_std": 0.07967609167098999, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14656248688697815, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "step": 3919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 573.0, "completions/mean_terminated_length": 573.0, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.784, "grad_norm": 0.17849113047122955, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46275870.0, "reward": 1.0750792026519775, "reward_std": 0.4555549919605255, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6501562595367432, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 3920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 517.5, "completions/mean_terminated_length": 517.5, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.7842, "grad_norm": 0.006089719478040934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46287958.0, "reward": 1.4908478260040283, "reward_std": 0.030908461660146713, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9816937446594238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9206249713897705, "step": 3921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 702.0, "completions/max_terminated_length": 702.0, "completions/mean_length": 349.4375, "completions/mean_terminated_length": 349.4375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.7844, "grad_norm": 0.2373572438955307, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46297757.0, "reward": 1.1942946910858154, "reward_std": 0.43891671299934387, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.701087474822998, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 3922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 512.3125, "completions/mean_terminated_length": 512.3125, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.7846, "grad_norm": 0.13842809200286865, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46311026.0, "reward": 1.2270103693008423, "reward_std": 0.43029314279556274, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.766518771648407, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.909375011920929, "step": 3923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 541.9375, "completions/mean_terminated_length": 509.8000183105469, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.7848, "grad_norm": 0.1687982827425003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46324017.0, "reward": 0.8609381914138794, "reward_std": 0.4555147588253021, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.534375011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6937499642372131, "step": 3924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 497.5625, "completions/mean_terminated_length": 497.5625, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.785, "grad_norm": 0.19009512662887573, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46336522.0, "reward": 1.3171323537826538, "reward_std": 0.3924977779388428, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8217625021934509, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "step": 3925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 773.625, "completions/mean_terminated_length": 715.84619140625, "completions/min_length": 547.0, "completions/min_terminated_length": 547.0, "epoch": 0.7852, "grad_norm": 0.0809774249792099, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46352788.0, "reward": 0.6270192861557007, "reward_std": 0.41029348969459534, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3790374994277954, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5299999713897705, "step": 3926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 882.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 633.75, "completions/mean_terminated_length": 633.75, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.7854, "grad_norm": 0.05100040137767792, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46367040.0, "reward": 1.001875400543213, "reward_std": 0.21325913071632385, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.7537499666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3125, "step": 3927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 639.6875, "completions/mean_terminated_length": 639.6875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.7856, "grad_norm": 0.0626349151134491, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46381643.0, "reward": 0.8902350068092346, "reward_std": 0.2388768047094345, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6554687023162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.590624988079071, "step": 3928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 922.5625, "completions/mean_terminated_length": 821.125, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.7858, "grad_norm": 0.19997526705265045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46400932.0, "reward": 0.40351587533950806, "reward_std": 0.43793144822120667, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3070312440395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.2593750059604645, "step": 3929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 299.0, "completions/max_terminated_length": 299.0, "completions/mean_length": 231.9375, "completions/mean_terminated_length": 231.9375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.786, "grad_norm": 0.2782233953475952, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46408659.0, "reward": 0.6566229462623596, "reward_std": 0.3295024335384369, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18824374675750732, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 3930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 481.9375, "completions/mean_terminated_length": 481.9375, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.7862, "grad_norm": 0.2036316692829132, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46422690.0, "reward": 1.0240135192871094, "reward_std": 0.4908491373062134, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5480250120162964, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 3931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 480.0, "completions/mean_terminated_length": 480.0, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.7864, "grad_norm": 0.09489211440086365, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46435058.0, "reward": 1.1928131580352783, "reward_std": 0.35740378499031067, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.8231250047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6937500238418579, "step": 3932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 423.5, "completions/mean_terminated_length": 423.5, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.7866, "grad_norm": 0.14834757149219513, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46445898.0, "reward": 1.2612509727478027, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7725000381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 309.0625, "completions/mean_terminated_length": 309.0625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.7868, "grad_norm": 0.25192397832870483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46454571.0, "reward": 1.3206915855407715, "reward_std": 0.38298988342285156, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8288812637329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 688.625, "completions/mean_terminated_length": 666.2667236328125, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.787, "grad_norm": 0.19337984919548035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46469253.0, "reward": 0.8094290494918823, "reward_std": 0.4989140033721924, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3688562512397766, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 3935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 321.5, "completions/mean_terminated_length": 321.5, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.7872, "grad_norm": 0.22691664099693298, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46478285.0, "reward": 0.7269415855407715, "reward_std": 0.3830983638763428, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26638123393058777, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 536.875, "completions/mean_terminated_length": 467.2857360839844, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.7874, "grad_norm": 0.20634503662586212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46490939.0, "reward": 1.1848444938659668, "reward_std": 0.5438317060470581, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8071874976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.731249988079071, "step": 3937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 952.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 633.375, "completions/mean_terminated_length": 633.375, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.7876, "grad_norm": 0.25362005829811096, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46505057.0, "reward": 1.0083853006362915, "reward_std": 0.5364247560501099, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.579268753528595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8756250143051147, "step": 3938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 530.8125, "completions/mean_terminated_length": 530.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.7878, "grad_norm": 0.08350222557783127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46517774.0, "reward": 0.7349225878715515, "reward_std": 0.1781713366508484, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4698437452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.653124988079071, "step": 3939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 616.75, "completions/mean_terminated_length": 589.6000366210938, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.788, "grad_norm": 0.06681715697050095, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46531866.0, "reward": 0.8984378576278687, "reward_std": 0.26025453209877014, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.859375, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.16875000298023224, "step": 3940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 498.1875, "completions/mean_terminated_length": 498.1875, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.7882, "grad_norm": 0.14927972853183746, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46543693.0, "reward": 0.8296101689338684, "reward_std": 0.3252643048763275, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5342187881469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 3941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 386.25, "completions/mean_terminated_length": 386.25, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.7884, "grad_norm": 0.1401572972536087, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46553809.0, "reward": 0.796250581741333, "reward_std": 0.21725216507911682, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5924999713897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5062500238418579, "step": 3942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 249.375, "completions/mean_terminated_length": 249.375, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.7886, "grad_norm": 0.0003572860441636294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46562087.0, "reward": 1.4987447261810303, "reward_std": 0.0011107606114819646, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.997487485408783, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 3943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 526.0, "completions/mean_terminated_length": 526.0, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.7888, "grad_norm": 0.2292625457048416, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46574727.0, "reward": 0.9285883903503418, "reward_std": 0.43139657378196716, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5446749925613403, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7987500429153442, "step": 3944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 622.3125, "completions/mean_terminated_length": 595.5333862304688, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.789, "grad_norm": 0.1872689574956894, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46588940.0, "reward": 0.9290509819984436, "reward_std": 0.5014858841896057, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5456000566482544, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8049999475479126, "step": 3945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 652.9375, "completions/mean_terminated_length": 652.9375, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.7892, "grad_norm": 0.14832334220409393, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46603835.0, "reward": 0.9687508940696716, "reward_std": 0.39801299571990967, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.625, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 3946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 335.25, "completions/mean_terminated_length": 335.25, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.7894, "grad_norm": 0.08393092453479767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46613023.0, "reward": 1.4389477968215942, "reward_std": 0.2373998910188675, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9403937458992004, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 3947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 294.875, "completions/mean_terminated_length": 294.875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.7896, "grad_norm": 0.2707761228084564, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46621533.0, "reward": 1.4394540786743164, "reward_std": 0.23752106726169586, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.94140625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 638.875, "completions/mean_terminated_length": 638.875, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.7898, "grad_norm": 0.09478268772363663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46635867.0, "reward": 1.1268630027770996, "reward_std": 0.33110255002975464, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.8787250518798828, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5299999713897705, "step": 3949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 363.0625, "completions/mean_terminated_length": 363.0625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.79, "grad_norm": 0.27933546900749207, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46645884.0, "reward": 0.7495946884155273, "reward_std": 0.4668487310409546, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31168749928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 680.5, "completions/mean_terminated_length": 566.0, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.7902, "grad_norm": 0.173700213432312, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46661076.0, "reward": 0.581316351890564, "reward_std": 0.4758439362049103, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2876312732696533, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6143749952316284, "step": 3951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 251.5625, "completions/mean_terminated_length": 251.5625, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.7904, "grad_norm": 0.16625842452049255, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46668733.0, "reward": 0.5820603966712952, "reward_std": 0.24506376683712006, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10161875188350677, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750667572021, "step": 3952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 416.375, "completions/mean_terminated_length": 416.375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.7906, "grad_norm": 0.36351823806762695, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46679747.0, "reward": 1.127307653427124, "reward_std": 0.4965737760066986, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6296125054359436, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9887499809265137, "step": 3953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 403.375, "completions/mean_terminated_length": 403.375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.7908, "grad_norm": 0.2245633602142334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46690265.0, "reward": 1.2344508171081543, "reward_std": 0.4003601670265198, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7813999652862549, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8837500214576721, "step": 3954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 428.625, "completions/mean_terminated_length": 428.625, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.791, "grad_norm": 0.25848329067230225, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46700979.0, "reward": 0.6935040950775146, "reward_std": 0.3343266248703003, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2620062530040741, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 3955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 627.1875, "completions/mean_terminated_length": 600.7333374023438, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.7912, "grad_norm": 0.18259857594966888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46714710.0, "reward": 0.9519164562225342, "reward_std": 0.5283092260360718, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5288312435150146, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8556250333786011, "step": 3956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 476.3125, "completions/mean_terminated_length": 476.3125, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.7914, "grad_norm": 0.17805390059947968, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46726603.0, "reward": 1.2028915882110596, "reward_std": 0.44923192262649536, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7182812690734863, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 3957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 456.5, "completions/mean_terminated_length": 456.5, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.7916, "grad_norm": 0.15431511402130127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46737779.0, "reward": 0.7862509489059448, "reward_std": 0.4248528778553009, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 457.4375, "completions/mean_terminated_length": 457.4375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.7918, "grad_norm": 0.1073409765958786, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46749290.0, "reward": 0.9274221658706665, "reward_std": 0.14488941431045532, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8548437356948853, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.24687498807907104, "step": 3959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 688.75, "completions/mean_terminated_length": 688.75, "completions/min_length": 536.0, "completions/min_terminated_length": 536.0, "epoch": 0.792, "grad_norm": 0.0500832237303257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46764982.0, "reward": 0.8564066886901855, "reward_std": 0.1457398533821106, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7128124833106995, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.46875, "step": 3960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 623.6875, "completions/mean_terminated_length": 623.6875, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.7922, "grad_norm": 0.23484279215335846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46778593.0, "reward": 0.9291259050369263, "reward_std": 0.4594358801841736, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4832499623298645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262499809265137, "step": 3961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 817.5625, "completions/mean_terminated_length": 788.0714721679688, "completions/min_length": 558.0, "completions/min_terminated_length": 558.0, "epoch": 0.7924, "grad_norm": 0.14628762006759644, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46795578.0, "reward": 0.8308758735656738, "reward_std": 0.42603006958961487, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5992506742477417, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5572937726974487, "step": 3962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 865.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 504.8125, "completions/mean_terminated_length": 504.8125, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.7926, "grad_norm": 0.22114740312099457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46807159.0, "reward": 0.8858602046966553, "reward_std": 0.43790188431739807, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45921874046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 3963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 406.5625, "completions/mean_terminated_length": 406.5625, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.7928, "grad_norm": 0.14954067766666412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46818048.0, "reward": 1.2449196577072144, "reward_std": 0.3610312342643738, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8648375272750854, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7837499976158142, "step": 3964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 556.25, "completions/mean_terminated_length": 343.6363830566406, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.793, "grad_norm": 0.23063908517360687, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46830452.0, "reward": 0.43648508191108704, "reward_std": 0.3839266896247864, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12296874821186066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.653124988079071, "step": 3965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 295.5625, "completions/mean_terminated_length": 295.5625, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.7932, "grad_norm": 0.00018092199752572924, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46838989.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 384.1875, "completions/mean_terminated_length": 384.1875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.7934, "grad_norm": 0.338364839553833, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46849584.0, "reward": 1.1425009965896606, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 433.8125, "completions/mean_terminated_length": 433.8125, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.7936, "grad_norm": 0.15280978381633759, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46860317.0, "reward": 0.786016583442688, "reward_std": 0.42443445324897766, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32203125953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 778.375, "completions/mean_terminated_length": 696.5, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.7938, "grad_norm": 0.14427267014980316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46877107.0, "reward": 0.672500491142273, "reward_std": 0.45923009514808655, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5325000286102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 3969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 293.0, "completions/max_terminated_length": 293.0, "completions/mean_length": 222.9375, "completions/mean_terminated_length": 222.9375, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.794, "grad_norm": 0.49832573533058167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46884738.0, "reward": 0.782082200050354, "reward_std": 0.42764812707901, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3141624927520752, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 3970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 318.5, "completions/mean_terminated_length": 318.5, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.7942, "grad_norm": 0.34817948937416077, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46894074.0, "reward": 0.7802229523658752, "reward_std": 0.42891180515289307, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3104437589645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 3971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 417.375, "completions/mean_terminated_length": 417.375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.7944, "grad_norm": 0.18535861372947693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46904288.0, "reward": 1.3074884414672852, "reward_std": 0.3372819423675537, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8649749755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8362499475479126, "step": 3972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 573.0, "completions/max_terminated_length": 573.0, "completions/mean_length": 363.6875, "completions/mean_terminated_length": 363.6875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.7946, "grad_norm": 0.3809475004673004, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46914139.0, "reward": 0.7297665476799011, "reward_std": 0.38169553875923157, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2720312476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 3973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 481.3125, "completions/mean_terminated_length": 445.13336181640625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.7948, "grad_norm": 0.12144594639539719, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46925760.0, "reward": 0.5925008654594421, "reward_std": 0.28516891598701477, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1849999874830246, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 3974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 442.4375, "completions/mean_terminated_length": 442.4375, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.795, "grad_norm": 0.39408883452415466, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46936583.0, "reward": 1.1964727640151978, "reward_std": 0.494223415851593, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7679437398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 3975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 483.9375, "completions/mean_terminated_length": 483.9375, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.7952, "grad_norm": 0.2239198237657547, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46948150.0, "reward": 0.8145319819450378, "reward_std": 0.31257855892181396, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5040625333786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.71875, "step": 3976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 324.5, "completions/mean_terminated_length": 324.5, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.7954, "grad_norm": 0.28724852204322815, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46958062.0, "reward": 1.3156447410583496, "reward_std": 0.39507704973220825, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8187875151634216, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "step": 3977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 670.25, "completions/mean_terminated_length": 619.7142944335938, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.7956, "grad_norm": 0.18051598966121674, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46973074.0, "reward": 1.0672659873962402, "reward_std": 0.4861747920513153, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.8220312595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.44062501192092896, "step": 3978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 408.375, "completions/mean_terminated_length": 408.375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.7958, "grad_norm": 0.3069533705711365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46983544.0, "reward": 1.167644739151001, "reward_std": 0.5229499936103821, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7102875709533691, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 364.0, "completions/max_terminated_length": 364.0, "completions/mean_length": 285.25, "completions/mean_terminated_length": 285.25, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.796, "grad_norm": 0.2059190422296524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46992044.0, "reward": 1.3283478021621704, "reward_std": 0.34095466136932373, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8441937565803528, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 3980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 447.625, "completions/mean_terminated_length": 447.625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.7962, "grad_norm": 0.1799066960811615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47003030.0, "reward": 0.6462509632110596, "reward_std": 0.2533104717731476, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22999998927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.893750011920929, "step": 3981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 445.1875, "completions/mean_terminated_length": 406.6000061035156, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.7964, "grad_norm": 0.3178591728210449, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47014601.0, "reward": 1.184128999710083, "reward_std": 0.4963725805282593, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7432562112808228, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8018749952316284, "step": 3982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 655.5, "completions/mean_terminated_length": 602.857177734375, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.7966, "grad_norm": 0.18363875150680542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47028913.0, "reward": 0.8339069485664368, "reward_std": 0.4900760054588318, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5428124666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.643750011920929, "step": 3983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 612.8125, "completions/mean_terminated_length": 554.0714721679688, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.7968, "grad_norm": 0.13260580599308014, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47042782.0, "reward": 0.54531329870224, "reward_std": 0.31715866923332214, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15312498807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.824999988079071, "step": 3984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 469.125, "completions/mean_terminated_length": 469.125, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.797, "grad_norm": 0.14819183945655823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47054112.0, "reward": 1.2595727443695068, "reward_std": 0.42821699380874634, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7691437602043152, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 3985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 896.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 674.625, "completions/mean_terminated_length": 674.625, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.7972, "grad_norm": 0.09634370356798172, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47069050.0, "reward": 0.6853100657463074, "reward_std": 0.18623939156532288, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3706187605857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7306250333786011, "step": 3986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 443.5, "completions/mean_terminated_length": 443.5, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.7974, "grad_norm": 0.34857288002967834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47080306.0, "reward": 1.2019414901733398, "reward_std": 0.45482364296913147, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7163812518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 3987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 592.4375, "completions/mean_terminated_length": 592.4375, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.7976, "grad_norm": 0.17892581224441528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47093865.0, "reward": 1.0799946784973145, "reward_std": 0.49036097526550293, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5974875092506409, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 3988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 600.25, "completions/mean_terminated_length": 572.0, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.7978, "grad_norm": 0.11194021254777908, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47108109.0, "reward": 0.6970193982124329, "reward_std": 0.31867313385009766, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.39403748512268066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6549999713897705, "step": 3989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 373.75, "completions/mean_terminated_length": 373.75, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.798, "grad_norm": 0.35024309158325195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47118233.0, "reward": 1.1451570987701416, "reward_std": 0.4264111816883087, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7278125286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 3990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 356.8125, "completions/mean_terminated_length": 356.8125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.7982, "grad_norm": 0.33952316641807556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47127590.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 638.0, "completions/mean_terminated_length": 638.0, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.7984, "grad_norm": 0.11239098757505417, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47141494.0, "reward": 0.9642196893692017, "reward_std": 0.4865482449531555, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4909375309944153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 3992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 348.1875, "completions/mean_terminated_length": 348.1875, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.7986, "grad_norm": 0.19562728703022003, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47151225.0, "reward": 1.2001415491104126, "reward_std": 0.4576508104801178, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7127813100814819, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 3993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 620.9375, "completions/mean_terminated_length": 563.357177734375, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.7988, "grad_norm": 0.09704845398664474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47165240.0, "reward": 0.5885944962501526, "reward_std": 0.33903396129608154, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2396875023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.75, "step": 3994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 348.25, "completions/mean_terminated_length": 348.25, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.799, "grad_norm": 0.08976389467716217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47174924.0, "reward": 1.4676446914672852, "reward_std": 0.1247052550315857, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9977874755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 3995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 859.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 721.9375, "completions/mean_terminated_length": 721.9375, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.7992, "grad_norm": 0.09013485163450241, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47190379.0, "reward": 0.7026574015617371, "reward_std": 0.3394039273262024, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2803124785423279, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 3996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 582.4375, "completions/mean_terminated_length": 582.4375, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.7994, "grad_norm": 0.38133496046066284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47203634.0, "reward": 0.9353790283203125, "reward_std": 0.5307338833808899, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49575626850128174, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 3997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 499.0, "completions/max_terminated_length": 499.0, "completions/mean_length": 407.6875, "completions/mean_terminated_length": 407.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.7996, "grad_norm": 0.4051739573478699, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47213965.0, "reward": 1.0862228870391846, "reward_std": 0.48365941643714905, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6099437475204468, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 3998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 469.125, "completions/mean_terminated_length": 432.13336181640625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.7998, "grad_norm": 0.09666785597801208, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47225743.0, "reward": 1.405832290649414, "reward_std": 0.3748890459537506, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9366625547409058, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9162500500679016, "step": 3999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 599.0, "completions/mean_terminated_length": 599.0, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.8, "grad_norm": 0.058409564197063446, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47239087.0, "reward": 0.6235948204994202, "reward_std": 0.24035851657390594, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18468749523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 4000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 754.125, "completions/mean_terminated_length": 736.1333618164062, "completions/min_length": 500.0, "completions/min_terminated_length": 500.0, "epoch": 0.8002, "grad_norm": 0.15491260588169098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47255457.0, "reward": 0.7780473828315735, "reward_std": 0.2495483011007309, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6185937523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.49687501788139343, "step": 4001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 500.25, "completions/mean_terminated_length": 500.25, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.8004, "grad_norm": 0.004394120071083307, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47267877.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 721.4375, "completions/mean_terminated_length": 701.2667236328125, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.8006, "grad_norm": 0.07064639031887054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47283372.0, "reward": 0.5106852650642395, "reward_std": 0.13758189976215363, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08386874198913574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943750262260437, "step": 4003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 524.1875, "completions/mean_terminated_length": 524.1875, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.8008, "grad_norm": 0.010515399277210236, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47296223.0, "reward": 0.5401260256767273, "reward_std": 0.023111840710043907, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08024999499320984, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 705.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 536.875, "completions/mean_terminated_length": 536.875, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.801, "grad_norm": 0.043797872960567474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47308669.0, "reward": 0.6410945653915405, "reward_std": 0.1340697854757309, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.28218749165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 4005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 543.4375, "completions/mean_terminated_length": 511.4000244140625, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.8012, "grad_norm": 0.2354750782251358, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47321572.0, "reward": 0.7542976140975952, "reward_std": 0.42414623498916626, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.38359373807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 4006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 347.375, "completions/mean_terminated_length": 347.375, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.8014, "grad_norm": 0.11776254326105118, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47331322.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 814.6875, "completions/mean_terminated_length": 689.1000366210938, "completions/min_length": 507.0, "completions/min_terminated_length": 507.0, "epoch": 0.8016, "grad_norm": 0.17061801254749298, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47348245.0, "reward": 0.5032817125320435, "reward_std": 0.4388855993747711, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3815625011920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.2874999940395355, "step": 4008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 419.6875, "completions/mean_terminated_length": 419.6875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.8018, "grad_norm": 0.24280868470668793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47358592.0, "reward": 0.8056259155273438, "reward_std": 0.32452821731567383, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.48624998331069946, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 4009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 855.0, "completions/mean_terminated_length": 798.6666870117188, "completions/min_length": 584.0, "completions/min_terminated_length": 584.0, "epoch": 0.802, "grad_norm": 0.23441989719867706, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47376608.0, "reward": 0.7539846897125244, "reward_std": 0.47307154536247253, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6954687833786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.19062501192092896, "step": 4010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 679.4375, "completions/mean_terminated_length": 656.4666748046875, "completions/min_length": 504.0, "completions/min_terminated_length": 504.0, "epoch": 0.8022, "grad_norm": 0.17503613233566284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47392023.0, "reward": 0.9906008243560791, "reward_std": 0.5417501926422119, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5437000393867493, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.89124995470047, "step": 4011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 370.625, "completions/mean_terminated_length": 370.625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.8024, "grad_norm": 0.3196094334125519, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47402513.0, "reward": 1.3189477920532227, "reward_std": 0.3869606852531433, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.825393795967102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 323.0, "completions/mean_terminated_length": 276.2666931152344, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.8026, "grad_norm": 0.3854161202907562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47411665.0, "reward": 1.3119289875030518, "reward_std": 0.4351998269557953, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.873856246471405, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8193750381469727, "step": 4013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 765.9375, "completions/mean_terminated_length": 729.0714721679688, "completions/min_length": 599.0, "completions/min_terminated_length": 599.0, "epoch": 0.8028, "grad_norm": 0.15018309652805328, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47428320.0, "reward": 0.8788789510726929, "reward_std": 0.5445901155471802, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.507756233215332, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.715624988079071, "step": 4014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 489.0, "completions/max_terminated_length": 489.0, "completions/mean_length": 379.0, "completions/mean_terminated_length": 379.0, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.803, "grad_norm": 0.4865962862968445, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47438448.0, "reward": 0.8456259369850159, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3787499964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 336.9375, "completions/mean_terminated_length": 336.9375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.8032, "grad_norm": 0.12342960387468338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47449215.0, "reward": 0.601460337638855, "reward_std": 0.23973597586154938, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14041873812675476, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 4016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 460.75, "completions/mean_terminated_length": 460.75, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.8034, "grad_norm": 0.13088282942771912, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47460251.0, "reward": 0.7166104316711426, "reward_std": 0.38847315311431885, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24571868777275085, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9605624675750732, "step": 4017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 599.8125, "completions/mean_terminated_length": 599.8125, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.8036, "grad_norm": 0.09971706569194794, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47474232.0, "reward": 0.6681259870529175, "reward_std": 0.2578928470611572, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2737500071525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8500000238418579, "step": 4018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 273.0, "completions/max_terminated_length": 273.0, "completions/mean_length": 228.375, "completions/mean_terminated_length": 228.375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.8038, "grad_norm": 0.23372389376163483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47481662.0, "reward": 1.2610164880752563, "reward_std": 0.4115479290485382, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7720312476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 425.375, "completions/mean_terminated_length": 425.375, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.804, "grad_norm": 0.21517783403396606, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47492052.0, "reward": 0.6260915398597717, "reward_std": 0.2469465583562851, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1896812468767166, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9243749976158142, "step": 4020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 348.8125, "completions/mean_terminated_length": 348.8125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.8042, "grad_norm": 0.10277116298675537, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47502145.0, "reward": 1.4674854278564453, "reward_std": 0.1259969174861908, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9974687695503235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9043749570846558, "step": 4021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 344.625, "completions/mean_terminated_length": 344.625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.8044, "grad_norm": 0.04885246977210045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47511739.0, "reward": 0.9780473113059998, "reward_std": 0.04120320454239845, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9560937881469727, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.140625, "step": 4022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 927.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 593.5, "completions/mean_terminated_length": 593.5, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8046, "grad_norm": 0.13855241239070892, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47525299.0, "reward": 1.356169581413269, "reward_std": 0.28731122612953186, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8998374938964844, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8087500333786011, "step": 4023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 277.8125, "completions/mean_terminated_length": 277.8125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.8048, "grad_norm": 0.39473894238471985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47533872.0, "reward": 0.8838289976119995, "reward_std": 0.43927648663520813, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45515620708465576, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8718749284744263, "step": 4024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 626.5625, "completions/mean_terminated_length": 600.0667114257812, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.805, "grad_norm": 0.1878274828195572, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47547977.0, "reward": 0.88453209400177, "reward_std": 0.4438134729862213, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5815625190734863, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.65625, "step": 4025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 233.625, "completions/mean_terminated_length": 233.625, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.8052, "grad_norm": 0.35715609788894653, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47555091.0, "reward": 1.2619197368621826, "reward_std": 0.4252519905567169, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7738375067710876, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 4026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 397.1875, "completions/mean_terminated_length": 397.1875, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.8054, "grad_norm": 0.24052362143993378, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47565702.0, "reward": 1.0111572742462158, "reward_std": 0.5041126608848572, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5223125219345093, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 4027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 433.0, "completions/mean_terminated_length": 433.0, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.8056, "grad_norm": 0.27241766452789307, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47576662.0, "reward": 1.0715103149414062, "reward_std": 0.5015627145767212, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5805187225341797, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9793750047683716, "step": 4028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 490.25, "completions/mean_terminated_length": 490.25, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.8058, "grad_norm": 0.13481734693050385, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47588714.0, "reward": 0.8048322200775146, "reward_std": 0.4093472361564636, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3596625030040741, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8799999952316284, "step": 4029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 460.25, "completions/mean_terminated_length": 460.25, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.806, "grad_norm": 0.16678379476070404, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47600222.0, "reward": 1.3803759813308716, "reward_std": 0.3246336877346039, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8857500553131104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 4030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 791.125, "completions/mean_terminated_length": 757.857177734375, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.8062, "grad_norm": 0.14674435555934906, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47618512.0, "reward": 0.6778099536895752, "reward_std": 0.32230129837989807, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.48061877489089966, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.49937498569488525, "step": 4031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 592.6875, "completions/mean_terminated_length": 563.933349609375, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.8064, "grad_norm": 0.13230933248996735, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47632267.0, "reward": 0.776250958442688, "reward_std": 0.4313901364803314, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.42750000953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 4032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 348.9375, "completions/mean_terminated_length": 348.9375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.8066, "grad_norm": 0.0004876307211816311, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47641418.0, "reward": 1.49851655960083, "reward_std": 0.0009375095251016319, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9970312118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 334.0, "completions/mean_terminated_length": 334.0, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.8068, "grad_norm": 0.3168897032737732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47650970.0, "reward": 1.1425665616989136, "reward_std": 0.4750525951385498, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6601312756538391, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 644.0, "completions/max_terminated_length": 644.0, "completions/mean_length": 503.8125, "completions/mean_terminated_length": 503.8125, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.807, "grad_norm": 0.2504200339317322, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47662743.0, "reward": 1.1425790786743164, "reward_std": 0.47506263852119446, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.66015625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 290.1875, "completions/mean_terminated_length": 290.1875, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.8072, "grad_norm": 0.31333595514297485, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47671450.0, "reward": 1.1507041454315186, "reward_std": 0.463295042514801, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6764062643051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1013.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 640.0625, "completions/mean_terminated_length": 640.0625, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.8074, "grad_norm": 0.11606907844543457, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47685483.0, "reward": 0.8571099638938904, "reward_std": 0.18516625463962555, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7142187356948853, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3968750238418579, "step": 4037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 566.75, "completions/mean_terminated_length": 501.4285888671875, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.8076, "grad_norm": 0.22629614174365997, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47698199.0, "reward": 0.8362508416175842, "reward_std": 0.5597184300422668, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42250001430511475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 4038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 947.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 460.125, "completions/mean_terminated_length": 460.125, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.8078, "grad_norm": 0.22565364837646484, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47709529.0, "reward": 1.0235165357589722, "reward_std": 0.48524171113967896, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5470312237739563, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 4039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 377.125, "completions/mean_terminated_length": 377.125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.808, "grad_norm": 0.4288724660873413, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47719371.0, "reward": 0.7862509489059448, "reward_std": 0.4248528778553009, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 347.125, "completions/mean_terminated_length": 347.125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.8082, "grad_norm": 0.20019754767417908, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47728765.0, "reward": 1.3798322677612305, "reward_std": 0.3244224190711975, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8846625089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 346.6875, "completions/mean_terminated_length": 346.6875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.8084, "grad_norm": 0.32103800773620605, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47738024.0, "reward": 0.617616593837738, "reward_std": 0.25026053190231323, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17273125052452087, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 4042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 374.5625, "completions/mean_terminated_length": 374.5625, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.8086, "grad_norm": 0.2044883668422699, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47748177.0, "reward": 1.2354695796966553, "reward_std": 0.4134766161441803, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7834374904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 4043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 219.0, "completions/max_terminated_length": 219.0, "completions/mean_length": 155.4375, "completions/mean_terminated_length": 155.4375, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "epoch": 0.8088, "grad_norm": 0.0004466203972697258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47754488.0, "reward": 1.499779224395752, "reward_std": 0.0004066909314133227, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.999556303024292, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "step": 4044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 949.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 568.4375, "completions/mean_terminated_length": 568.4375, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.809, "grad_norm": 0.19296573102474213, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47767663.0, "reward": 0.9870319366455078, "reward_std": 0.38996368646621704, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6615625023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.71875, "step": 4045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 578.5, "completions/mean_terminated_length": 578.5, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.8092, "grad_norm": 0.060613323003053665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47780999.0, "reward": 0.7110915184020996, "reward_std": 0.22956952452659607, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.42218124866485596, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6243749856948853, "step": 4046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 470.5625, "completions/mean_terminated_length": 470.5625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.8094, "grad_norm": 0.1506878137588501, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47792320.0, "reward": 1.175622820854187, "reward_std": 0.44411182403564453, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7262437343597412, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8931249976158142, "step": 4047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 415.75, "completions/mean_terminated_length": 415.75, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.8096, "grad_norm": 0.2563946545124054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47802876.0, "reward": 0.9903134107589722, "reward_std": 0.469145804643631, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5431249737739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 4048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 311.5625, "completions/mean_terminated_length": 311.5625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.8098, "grad_norm": 0.28199657797813416, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47811509.0, "reward": 1.3186416625976562, "reward_std": 0.38847607374191284, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8247812986373901, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 4049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 497.1875, "completions/mean_terminated_length": 497.1875, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.81, "grad_norm": 0.13922913372516632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47823576.0, "reward": 1.318904161453247, "reward_std": 0.3886060416698456, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8253062963485718, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 4050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 402.875, "completions/mean_terminated_length": 402.875, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.8102, "grad_norm": 0.35336679220199585, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47833654.0, "reward": 1.1901414394378662, "reward_std": 0.46565762162208557, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8177812695503235, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7543749809265137, "step": 4051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 551.875, "completions/mean_terminated_length": 551.875, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.8104, "grad_norm": 0.3474694788455963, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47846740.0, "reward": 1.1050446033477783, "reward_std": 0.473646879196167, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6475875377655029, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 4052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 404.875, "completions/mean_terminated_length": 404.875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.8106, "grad_norm": 0.1296946108341217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47857250.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 720.5, "completions/mean_terminated_length": 700.2667236328125, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.8108, "grad_norm": 0.08217090368270874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47873914.0, "reward": 0.5742946863174438, "reward_std": 0.28246915340423584, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14858749508857727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 4054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 570.1875, "completions/mean_terminated_length": 570.1875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.811, "grad_norm": 0.1109635978937149, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47886877.0, "reward": 0.8382036685943604, "reward_std": 0.2500268816947937, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6139062643051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5718749761581421, "step": 4055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 851.5625, "completions/mean_terminated_length": 773.1818237304688, "completions/min_length": 599.0, "completions/min_terminated_length": 599.0, "epoch": 0.8112, "grad_norm": 0.24598225951194763, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47905254.0, "reward": 0.8242850303649902, "reward_std": 0.6708451509475708, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.4610687494277954, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5643749833106995, "step": 4056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 293.875, "completions/mean_terminated_length": 293.875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.8114, "grad_norm": 0.41494235396385193, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47913348.0, "reward": 1.3208322525024414, "reward_std": 0.3830597996711731, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291624784469604, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 275.125, "completions/mean_terminated_length": 275.125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.8116, "grad_norm": 0.0005195480771362782, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47922902.0, "reward": 1.4993791580200195, "reward_std": 0.0005750601994805038, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987562894821167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 4058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 402.4375, "completions/mean_terminated_length": 402.4375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.8118, "grad_norm": 0.2583375871181488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47933613.0, "reward": 1.0938165187835693, "reward_std": 0.4760848581790924, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6251312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9393749833106995, "step": 4059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 804.5, "completions/mean_terminated_length": 704.727294921875, "completions/min_length": 556.0, "completions/min_terminated_length": 556.0, "epoch": 0.812, "grad_norm": 0.2825433015823364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47950949.0, "reward": 0.5782034397125244, "reward_std": 0.43274930119514465, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4689062535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 4060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 560.25, "completions/mean_terminated_length": 529.3333740234375, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.8122, "grad_norm": 0.1349383145570755, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47964201.0, "reward": 0.6265384554862976, "reward_std": 0.3663664758205414, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1905750036239624, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8974999785423279, "step": 4061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 350.125, "completions/mean_terminated_length": 350.125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.8124, "grad_norm": 0.19951677322387695, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47973851.0, "reward": 1.3800948858261108, "reward_std": 0.32452547550201416, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851875066757202, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 580.25, "completions/mean_terminated_length": 580.25, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.8126, "grad_norm": 0.09973850101232529, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47986943.0, "reward": 0.8136695027351379, "reward_std": 0.3535647988319397, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4398375153541565, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8025000095367432, "step": 4063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 511.5625, "completions/mean_terminated_length": 511.5625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.8128, "grad_norm": 0.2167036235332489, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47998888.0, "reward": 0.7519291043281555, "reward_std": 0.380763977766037, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31635624170303345, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.919374942779541, "step": 4064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 474.0625, "completions/mean_terminated_length": 474.0625, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.813, "grad_norm": 0.06786274164915085, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48010809.0, "reward": 0.8065599203109741, "reward_std": 0.2206873595714569, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6131187677383423, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4868749976158142, "step": 4065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 525.8125, "completions/mean_terminated_length": 525.8125, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.8132, "grad_norm": 0.2174772024154663, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48022998.0, "reward": 1.162969708442688, "reward_std": 0.4544287323951721, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7009375095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 4066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 622.75, "completions/mean_terminated_length": 596.0000610351562, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.8134, "grad_norm": 0.18779781460762024, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48036898.0, "reward": 0.844594419002533, "reward_std": 0.41346824169158936, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5641875267028809, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6200000047683716, "step": 4067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 597.3125, "completions/mean_terminated_length": 568.86669921875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.8136, "grad_norm": 0.15693385899066925, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48050359.0, "reward": 1.2269539833068848, "reward_std": 0.501449704170227, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7664062976837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 4068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 313.25, "completions/mean_terminated_length": 313.25, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.8138, "grad_norm": 0.15774358808994293, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48059307.0, "reward": 1.4394540786743164, "reward_std": 0.23752105236053467, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.94140625, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 465.5625, "completions/mean_terminated_length": 465.5625, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.814, "grad_norm": 0.19627782702445984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48070660.0, "reward": 1.2168760299682617, "reward_std": 0.433899849653244, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7462500333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.925000011920929, "step": 4070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 649.3125, "completions/mean_terminated_length": 649.3125, "completions/min_length": 455.0, "completions/min_terminated_length": 455.0, "epoch": 0.8142, "grad_norm": 0.14575247466564178, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48085353.0, "reward": 1.163985252380371, "reward_std": 0.4478721022605896, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7029687762260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 1024.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 928.9375, "completions/mean_terminated_length": 806.7142944335938, "completions/min_length": 639.0, "completions/min_terminated_length": 639.0, "epoch": 0.8144, "grad_norm": 0.22591425478458405, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48104008.0, "reward": 0.3864066004753113, "reward_std": 0.5288940668106079, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21031250059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.3812499940395355, "step": 4072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 726.875, "completions/mean_terminated_length": 658.3077392578125, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.8146, "grad_norm": 0.13298358023166656, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48120038.0, "reward": 0.7345319390296936, "reward_std": 0.5334708094596863, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4065625071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6875, "step": 4073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 751.0625, "completions/mean_terminated_length": 712.0714721679688, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.8148, "grad_norm": 0.11475566029548645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48137783.0, "reward": 0.7797659635543823, "reward_std": 0.35034066438674927, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6845312714576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.27812498807907104, "step": 4074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 463.375, "completions/mean_terminated_length": 463.375, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.815, "grad_norm": 0.00672491081058979, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48150845.0, "reward": 0.5498322248458862, "reward_std": 0.014043881557881832, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09966249763965607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 562.3125, "completions/mean_terminated_length": 531.5333862304688, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.8152, "grad_norm": 0.2518395781517029, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48164178.0, "reward": 1.085469365119934, "reward_std": 0.4770191013813019, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7334375381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 608.6875, "completions/mean_terminated_length": 608.6875, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.8154, "grad_norm": 0.15243099629878998, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48178461.0, "reward": 0.8954945802688599, "reward_std": 0.4442918300628662, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.478487491607666, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8662500381469727, "step": 4077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 376.375, "completions/mean_terminated_length": 376.375, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.8156, "grad_norm": 0.14900122582912445, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48188163.0, "reward": 0.6706258058547974, "reward_std": 0.2522943913936615, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2787500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.856249988079071, "step": 4078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 545.5625, "completions/mean_terminated_length": 545.5625, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.8158, "grad_norm": 0.06308719515800476, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48200796.0, "reward": 0.5855477452278137, "reward_std": 0.08867644518613815, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17109373211860657, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 445.0, "completions/max_terminated_length": 445.0, "completions/mean_length": 387.4375, "completions/mean_terminated_length": 387.4375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.816, "grad_norm": 0.22905661165714264, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48210947.0, "reward": 1.380403995513916, "reward_std": 0.324644535779953, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8858062624931335, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 884.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 560.9375, "completions/mean_terminated_length": 560.9375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.8162, "grad_norm": 0.16906622052192688, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48223986.0, "reward": 0.8527196049690247, "reward_std": 0.4013260006904602, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.45543748140335083, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8574999570846558, "step": 4081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 520.0, "completions/mean_terminated_length": 520.0, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.8164, "grad_norm": 0.008421688340604305, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48238210.0, "reward": 0.5402821898460388, "reward_std": 0.01536132674664259, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08056249469518661, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 344.875, "completions/mean_terminated_length": 344.875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.8166, "grad_norm": 0.28374364972114563, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48247936.0, "reward": 0.9001508951187134, "reward_std": 0.47904762625694275, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4253000020980835, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 4083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 929.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 743.125, "completions/mean_terminated_length": 743.125, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.8168, "grad_norm": 0.06902765482664108, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48263986.0, "reward": 0.6320852041244507, "reward_std": 0.17476767301559448, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26416873931884766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8006249666213989, "step": 4084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 900.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 620.4375, "completions/mean_terminated_length": 620.4375, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.817, "grad_norm": 0.19616448879241943, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48277961.0, "reward": 1.0292977094650269, "reward_std": 0.48458659648895264, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.55859375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "step": 4085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 443.75, "completions/mean_terminated_length": 309.8461608886719, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.8172, "grad_norm": 0.40987950563430786, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48289237.0, "reward": 0.7343603372573853, "reward_std": 0.5706470012664795, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3437187373638153, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 4086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 614.4375, "completions/mean_terminated_length": 587.1333618164062, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.8174, "grad_norm": 0.2122814804315567, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48303036.0, "reward": 0.7131634950637817, "reward_std": 0.43116042017936707, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3013250231742859, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8487499952316284, "step": 4087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 355.1875, "completions/mean_terminated_length": 355.1875, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.8176, "grad_norm": 0.3652482330799103, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48312687.0, "reward": 0.8446633815765381, "reward_std": 0.4552052319049835, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3768250048160553, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 4088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 786.4375, "completions/mean_terminated_length": 731.6154174804688, "completions/min_length": 520.0, "completions/min_terminated_length": 520.0, "epoch": 0.8178, "grad_norm": 0.12736214697360992, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48329542.0, "reward": 0.5210164189338684, "reward_std": 0.3500717580318451, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16703124344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7531249523162842, "step": 4089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 605.5625, "completions/mean_terminated_length": 605.5625, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.818, "grad_norm": 0.0465925857424736, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48343279.0, "reward": 0.5632448196411133, "reward_std": 0.08886124193668365, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12648749351501465, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9275000095367432, "step": 4090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 391.375, "completions/mean_terminated_length": 391.375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.8182, "grad_norm": 0.08401583135128021, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48353605.0, "reward": 1.4426167011260986, "reward_std": 0.22603148221969604, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9477312564849854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 943.9375, "completions/mean_terminated_length": 767.7999877929688, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.8184, "grad_norm": 0.2122567892074585, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48373508.0, "reward": 0.2338283658027649, "reward_std": 0.3853655457496643, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09265625476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.18437498807907104, "step": 4092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 674.25, "completions/mean_terminated_length": 674.25, "completions/min_length": 463.0, "completions/min_terminated_length": 463.0, "epoch": 0.8186, "grad_norm": 0.25455451011657715, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48388120.0, "reward": 1.021628975868225, "reward_std": 0.4462983310222626, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6057562232017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8768749833106995, "step": 4093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 474.8125, "completions/mean_terminated_length": 474.8125, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.8188, "grad_norm": 0.1990136206150055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48399669.0, "reward": 1.137876033782959, "reward_std": 0.4817858040332794, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6507500410079956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "step": 4094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1019.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 457.5, "completions/mean_terminated_length": 457.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.819, "grad_norm": 0.25891613960266113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48411261.0, "reward": 0.7061601281166077, "reward_std": 0.19011762738227844, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.41231873631477356, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7018749713897705, "step": 4095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 700.0, "completions/mean_terminated_length": 653.7142944335938, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.8192, "grad_norm": 0.3510224223136902, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48426397.0, "reward": 0.8378129601478577, "reward_std": 0.4195001423358917, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6756249666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3687499761581421, "step": 4096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 502.875, "completions/mean_terminated_length": 502.875, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.8194, "grad_norm": 0.23489733040332794, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48438491.0, "reward": 1.0067195892333984, "reward_std": 0.4580174684524536, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5759375095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 4097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 595.5, "completions/mean_terminated_length": 566.933349609375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.8196, "grad_norm": 0.12570270895957947, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48453715.0, "reward": 0.6917194128036499, "reward_std": 0.2814892828464508, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.44593751430511475, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.550000011920929, "step": 4098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 779.3125, "completions/mean_terminated_length": 763.0000610351562, "completions/min_length": 526.0, "completions/min_terminated_length": 526.0, "epoch": 0.8198, "grad_norm": 0.13763128221035004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48470456.0, "reward": 0.751094400882721, "reward_std": 0.28690463304519653, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5646874904632568, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4625000059604645, "step": 4099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 222.0, "completions/max_terminated_length": 222.0, "completions/mean_length": 187.9375, "completions/mean_terminated_length": 187.9375, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.82, "grad_norm": 0.0008125014719553292, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48477191.0, "reward": 1.4991728067398071, "reward_std": 0.0005645328783430159, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983437657356262, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 419.8125, "completions/mean_terminated_length": 419.8125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.8202, "grad_norm": 0.21604210138320923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48488132.0, "reward": 0.7123916149139404, "reward_std": 0.3907323479652405, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23728126287460327, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 4101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 423.0625, "completions/mean_terminated_length": 423.0625, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.8204, "grad_norm": 0.005228773690760136, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48498917.0, "reward": 1.4966415166854858, "reward_std": 0.00749615952372551, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9932812452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 407.0, "completions/max_terminated_length": 407.0, "completions/mean_length": 320.375, "completions/mean_terminated_length": 320.375, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.8206, "grad_norm": 0.3233921229839325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48507931.0, "reward": 1.320822834968567, "reward_std": 0.3830551207065582, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.829143762588501, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 819.4375, "completions/mean_terminated_length": 751.25, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.8208, "grad_norm": 0.14646948873996735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48525698.0, "reward": 0.463034987449646, "reward_std": 0.3111237585544586, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1760687530040741, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6206250190734863, "step": 4104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 511.375, "completions/mean_terminated_length": 511.375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.821, "grad_norm": 0.3582397699356079, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48538088.0, "reward": 1.2768759727478027, "reward_std": 0.3997112512588501, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8037499785423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 4105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 468.5, "completions/mean_terminated_length": 468.5, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.8212, "grad_norm": 0.1789720356464386, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48550032.0, "reward": 1.022679090499878, "reward_std": 0.4925462305545807, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.545356273651123, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 401.9375, "completions/mean_terminated_length": 401.9375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.8214, "grad_norm": 0.2570357322692871, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48560591.0, "reward": 1.1406915187835693, "reward_std": 0.4774612486362457, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6563812494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 390.625, "completions/mean_terminated_length": 390.625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.8216, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48570745.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 483.625, "completions/mean_terminated_length": 447.60003662109375, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.8218, "grad_norm": 0.1546885222196579, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48582707.0, "reward": 0.7213256359100342, "reward_std": 0.33203205466270447, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4426500201225281, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6399999856948853, "step": 4109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 876.75, "completions/mean_terminated_length": 827.6666870117188, "completions/min_length": 707.0, "completions/min_terminated_length": 707.0, "epoch": 0.822, "grad_norm": 0.1429947316646576, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48600815.0, "reward": 0.628281831741333, "reward_std": 0.45376500487327576, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4440625309944153, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.40625, "step": 4110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 524.0, "completions/mean_length": 453.1875, "completions/mean_terminated_length": 415.13336181640625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.8222, "grad_norm": 0.3241731524467468, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48611986.0, "reward": 0.9809353947639465, "reward_std": 0.5500818490982056, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5243687629699707, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 4111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 462.375, "completions/mean_terminated_length": 462.375, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.8224, "grad_norm": 0.086991086602211, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48623560.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 530.4375, "completions/mean_terminated_length": 530.4375, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.8226, "grad_norm": 0.1577443927526474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48636431.0, "reward": 0.6675009727478027, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20999999344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 344.8125, "completions/mean_terminated_length": 344.8125, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.8228, "grad_norm": 0.14069288969039917, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48646060.0, "reward": 1.4112385511398315, "reward_std": 0.25073006749153137, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9474749565124512, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 4114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 763.0625, "completions/mean_terminated_length": 676.0833740234375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.823, "grad_norm": 0.18809452652931213, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48662381.0, "reward": 0.45984435081481934, "reward_std": 0.30371230840682983, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16968747973442078, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 4115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 770.1875, "completions/mean_terminated_length": 685.5833740234375, "completions/min_length": 530.0, "completions/min_terminated_length": 530.0, "epoch": 0.8232, "grad_norm": 0.11354878544807434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48678448.0, "reward": 0.4273444414138794, "reward_std": 0.25991740822792053, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10468749701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6937499642372131, "step": 4116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 586.3125, "completions/mean_terminated_length": 557.1333618164062, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.8234, "grad_norm": 0.17607958614826202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48691557.0, "reward": 0.7182820439338684, "reward_std": 0.37936288118362427, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3740624785423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.762499988079071, "step": 4117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 567.0625, "completions/mean_terminated_length": 536.6000366210938, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.8236, "grad_norm": 0.2648114860057831, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48704502.0, "reward": 1.0096882581710815, "reward_std": 0.488415002822876, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6443749666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "step": 4118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 426.0625, "completions/mean_terminated_length": 426.0625, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.8238, "grad_norm": 0.12869124114513397, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48715239.0, "reward": 0.6515634059906006, "reward_std": 0.2485060840845108, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24062499403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 4119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 576.125, "completions/mean_terminated_length": 576.125, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.824, "grad_norm": 0.21643809974193573, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48729113.0, "reward": 0.9645290374755859, "reward_std": 0.43996283411979675, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5540562868118286, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8493750095367432, "step": 4120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 369.5625, "completions/mean_terminated_length": 369.5625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.8242, "grad_norm": 0.35654258728027344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48739234.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 603.125, "completions/mean_terminated_length": 603.125, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.8244, "grad_norm": 0.007821205072104931, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48752484.0, "reward": 0.5478728413581848, "reward_std": 0.023259053006768227, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09574374556541443, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 4122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 389.75, "completions/mean_terminated_length": 389.75, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.8246, "grad_norm": 0.21166834235191345, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48762656.0, "reward": 1.2280353307724, "reward_std": 0.4279422461986542, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.768568754196167, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9018750190734863, "step": 4123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 471.3125, "completions/mean_terminated_length": 471.3125, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.8248, "grad_norm": 0.1596006602048874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48773957.0, "reward": 1.0843510627746582, "reward_std": 0.4857964515686035, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6062000393867493, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 254.8125, "completions/mean_terminated_length": 254.8125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.825, "grad_norm": 0.5160886645317078, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48781698.0, "reward": 1.3207571506500244, "reward_std": 0.3830225467681885, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290125131607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 934.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 616.9375, "completions/mean_terminated_length": 616.9375, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.8252, "grad_norm": 0.06666310876607895, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48795121.0, "reward": 0.6794538497924805, "reward_std": 0.1445753127336502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.35890626907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.778124988079071, "step": 4126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 372.75, "completions/mean_terminated_length": 372.75, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.8254, "grad_norm": 0.1539764702320099, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48804669.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 396.0625, "completions/mean_terminated_length": 354.20001220703125, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.8256, "grad_norm": 0.40169569849967957, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48814558.0, "reward": 1.227004051208496, "reward_std": 0.5014799237251282, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7665062546730042, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 332.0, "completions/max_terminated_length": 332.0, "completions/mean_length": 254.25, "completions/mean_terminated_length": 254.25, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.8258, "grad_norm": 0.4186055064201355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48822578.0, "reward": 1.3773009777069092, "reward_std": 0.33306699991226196, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8796000480651855, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 275.25, "completions/mean_terminated_length": 275.25, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.826, "grad_norm": 0.4059566259384155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48831174.0, "reward": 1.2048321962356567, "reward_std": 0.45052218437194824, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7221624851226807, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 429.0, "completions/mean_terminated_length": 429.0, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.8262, "grad_norm": 0.19466470181941986, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48843766.0, "reward": 1.2653132677078247, "reward_std": 0.35260629653930664, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8431249856948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.800000011920929, "step": 4131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 530.1875, "completions/mean_terminated_length": 530.1875, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.8264, "grad_norm": 0.20319275557994843, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48856153.0, "reward": 0.8547664284706116, "reward_std": 0.41817086935043335, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3970312476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656250238418579, "step": 4132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 531.4375, "completions/mean_terminated_length": 531.4375, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.8266, "grad_norm": 0.08117437362670898, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48868384.0, "reward": 1.3907570838928223, "reward_std": 0.2971307337284088, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9065124988555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9412499666213989, "step": 4133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 651.1875, "completions/mean_terminated_length": 651.1875, "completions/min_length": 476.0, "completions/min_terminated_length": 476.0, "epoch": 0.8268, "grad_norm": 0.08177727460861206, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48882739.0, "reward": 0.7170730233192444, "reward_std": 0.388769268989563, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24664373695850372, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 751.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 577.6875, "completions/mean_terminated_length": 577.6875, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.827, "grad_norm": 0.14664149284362793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48895966.0, "reward": 0.9595979452133179, "reward_std": 0.4914175868034363, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48169374465942383, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 695.6875, "completions/mean_terminated_length": 673.800048828125, "completions/min_length": 436.0, "completions/min_terminated_length": 436.0, "epoch": 0.8272, "grad_norm": 0.15874500572681427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48910985.0, "reward": 0.7178133726119995, "reward_std": 0.4185704290866852, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.31062501668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8562499284744263, "step": 4136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 688.4375, "completions/mean_terminated_length": 640.5, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.8274, "grad_norm": 0.2911577820777893, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48925584.0, "reward": 0.8431257605552673, "reward_std": 0.5425624251365662, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4362499713897705, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 4137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1006.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 618.875, "completions/mean_terminated_length": 618.875, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.8276, "grad_norm": 0.16537053883075714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48939710.0, "reward": 0.8957697153091431, "reward_std": 0.4331907629966736, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.47903749346733093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8862500190734863, "step": 4138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 553.625, "completions/mean_terminated_length": 553.625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.8278, "grad_norm": 0.07349240779876709, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48953896.0, "reward": 0.605079174041748, "reward_std": 0.2386225014925003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14765624701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 581.5, "completions/mean_terminated_length": 518.2857666015625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.828, "grad_norm": 0.192828968167305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48967232.0, "reward": 1.1131131649017334, "reward_std": 0.5328918695449829, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7887250185012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6612499952316284, "step": 4140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 397.5, "completions/mean_terminated_length": 397.5, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.8282, "grad_norm": 0.21104887127876282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48977160.0, "reward": 1.2099196910858154, "reward_std": 0.4431396722793579, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7323375344276428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "step": 4141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 328.0, "completions/max_terminated_length": 328.0, "completions/mean_length": 268.75, "completions/mean_terminated_length": 268.75, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.8284, "grad_norm": 0.4699373245239258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48985044.0, "reward": 1.1964261531829834, "reward_std": 0.465208500623703, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7053499817848206, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.981249988079071, "step": 4142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 505.9375, "completions/mean_terminated_length": 505.9375, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.8286, "grad_norm": 0.17196699976921082, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48997299.0, "reward": 0.9236477613449097, "reward_std": 0.4672979712486267, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47229376435279846, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256249666213989, "step": 4143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 580.0625, "completions/mean_terminated_length": 580.0625, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.8288, "grad_norm": 0.25125551223754883, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49010260.0, "reward": 1.190845251083374, "reward_std": 0.47299328446388245, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6941875219345093, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "step": 4144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 421.6875, "completions/mean_terminated_length": 381.5333557128906, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.829, "grad_norm": 0.3159506022930145, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49021215.0, "reward": 1.2551414966583252, "reward_std": 0.472969114780426, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8227812647819519, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8481249809265137, "step": 4145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 379.0, "completions/max_terminated_length": 379.0, "completions/mean_length": 318.125, "completions/mean_terminated_length": 318.125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.8292, "grad_norm": 0.0003748243616428226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49029969.0, "reward": 1.499435305595398, "reward_std": 0.0005503871943801641, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9988687634468079, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 4146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 652.4375, "completions/mean_terminated_length": 652.4375, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.8294, "grad_norm": 0.20377281308174133, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49044520.0, "reward": 0.9162479043006897, "reward_std": 0.46748700737953186, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.45749375224113464, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 4147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 971.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 688.25, "completions/mean_terminated_length": 688.25, "completions/min_length": 529.0, "completions/min_terminated_length": 529.0, "epoch": 0.8296, "grad_norm": 0.10003269463777542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49060092.0, "reward": 0.6684384346008301, "reward_std": 0.24274097383022308, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2743750214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 4148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 668.9375, "completions/mean_terminated_length": 645.2667236328125, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.8298, "grad_norm": 0.1659526526927948, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49074843.0, "reward": 0.8734384179115295, "reward_std": 0.5181236863136292, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43437498807907104, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8874999284744263, "step": 4149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 605.125, "completions/mean_terminated_length": 605.125, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.83, "grad_norm": 0.09702199697494507, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49088349.0, "reward": 0.6063165664672852, "reward_std": 0.23809203505516052, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1501312404870987, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 988.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 536.125, "completions/mean_terminated_length": 536.125, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.8302, "grad_norm": 0.3263857364654541, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49100735.0, "reward": 1.0487384796142578, "reward_std": 0.47791826725006104, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.597474992275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9050000309944153, "step": 4151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 590.0, "completions/mean_terminated_length": 590.0, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.8304, "grad_norm": 0.17883923649787903, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49113807.0, "reward": 0.815535306930542, "reward_std": 0.42251792550086975, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38106876611709595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 4152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 299.3125, "completions/mean_terminated_length": 299.3125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.8306, "grad_norm": 0.27010077238082886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49122068.0, "reward": 1.4373197555541992, "reward_std": 0.2472989559173584, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9371374845504761, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 4153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 268.625, "completions/mean_terminated_length": 268.625, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.8308, "grad_norm": 0.4288650453090668, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49130158.0, "reward": 1.3194822072982788, "reward_std": 0.38722601532936096, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8264625072479248, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "step": 4154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 684.9375, "completions/mean_terminated_length": 684.9375, "completions/min_length": 495.0, "completions/min_terminated_length": 495.0, "epoch": 0.831, "grad_norm": 0.08847825229167938, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49145613.0, "reward": 0.749297559261322, "reward_std": 0.19271264970302582, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4985937178134918, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.628125011920929, "step": 4155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 575.0, "completions/mean_terminated_length": 575.0, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.8312, "grad_norm": 0.18508358299732208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49158877.0, "reward": 1.026250958442688, "reward_std": 0.48795098066329956, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5525000095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 514.125, "completions/mean_terminated_length": 514.125, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.8314, "grad_norm": 0.1464904248714447, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49171311.0, "reward": 0.6980476379394531, "reward_std": 0.2650136351585388, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3335937559604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.809374988079071, "step": 4157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 356.375, "completions/mean_terminated_length": 356.375, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.8316, "grad_norm": 0.21319237351417542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49181141.0, "reward": 1.3184071779251099, "reward_std": 0.38835954666137695, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8243125677108765, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 509.1875, "completions/mean_terminated_length": 474.86669921875, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.8318, "grad_norm": 0.22051887214183807, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49193384.0, "reward": 0.9892977476119995, "reward_std": 0.5422846674919128, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5410937666893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 4159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 532.3125, "completions/mean_terminated_length": 532.3125, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.832, "grad_norm": 0.09697235375642776, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49205693.0, "reward": 0.7167195081710815, "reward_std": 0.19580768048763275, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4334374964237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6875, "step": 4160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 631.6875, "completions/mean_terminated_length": 541.1538696289062, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.8322, "grad_norm": 0.1416465789079666, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49219608.0, "reward": 0.6253129839897156, "reward_std": 0.34594371914863586, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.43812501430511475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49375003576278687, "step": 4161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 632.0625, "completions/mean_terminated_length": 605.933349609375, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.8324, "grad_norm": 0.03457082808017731, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49233977.0, "reward": 0.5324196219444275, "reward_std": 0.1642770767211914, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1273374855518341, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8650000095367432, "step": 4162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 358.375, "completions/mean_terminated_length": 358.375, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.8326, "grad_norm": 0.12740616500377655, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49245503.0, "reward": 0.9095316529273987, "reward_std": 0.1423080563545227, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8190624713897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3187499940395355, "step": 4163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 275.875, "completions/mean_terminated_length": 275.875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.8328, "grad_norm": 0.37575408816337585, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49253533.0, "reward": 0.9974666833877563, "reward_std": 0.47396573424339294, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5574312210083008, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8880625367164612, "step": 4164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 381.375, "completions/mean_terminated_length": 381.375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.833, "grad_norm": 0.17153196036815643, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49264531.0, "reward": 1.439507246017456, "reward_std": 0.2375352531671524, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9415124654769897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 481.5625, "completions/mean_terminated_length": 481.5625, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.8332, "grad_norm": 0.3424701690673828, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49276428.0, "reward": 1.2568385601043701, "reward_std": 0.43393591046333313, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7636750340461731, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 4166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 555.5, "completions/mean_terminated_length": 555.5, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.8334, "grad_norm": 0.136177659034729, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49289460.0, "reward": 0.8034350872039795, "reward_std": 0.22583109140396118, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6068687438964844, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4493749737739563, "step": 4167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 451.9375, "completions/mean_terminated_length": 451.9375, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.8336, "grad_norm": 0.10918905586004257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49300979.0, "reward": 0.6696634292602539, "reward_std": 0.32424652576446533, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21432499587535858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "step": 4168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 415.0625, "completions/mean_terminated_length": 415.0625, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.8338, "grad_norm": 0.22358019649982452, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49311572.0, "reward": 0.9050009846687317, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4350000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 483.875, "completions/mean_terminated_length": 483.875, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.834, "grad_norm": 0.20306923985481262, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49323522.0, "reward": 0.8150787353515625, "reward_std": 0.22132377326488495, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6301562786102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.47187501192092896, "step": 4170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 371.4375, "completions/mean_terminated_length": 371.4375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.8342, "grad_norm": 0.21735544502735138, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49333881.0, "reward": 0.6008884310722351, "reward_std": 0.2397719919681549, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13927499949932098, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 788.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 440.8125, "completions/mean_terminated_length": 440.8125, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.8344, "grad_norm": 0.20262926816940308, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49345334.0, "reward": 1.4078915119171143, "reward_std": 0.26130160689353943, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9407812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 538.8125, "completions/mean_terminated_length": 538.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.8346, "grad_norm": 0.2634097933769226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49357971.0, "reward": 1.1927638053894043, "reward_std": 0.4701598286628723, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6980249881744385, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.981249988079071, "step": 4173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 580.5625, "completions/mean_terminated_length": 517.2142944335938, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.8348, "grad_norm": 0.21333090960979462, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49371420.0, "reward": 0.7188289165496826, "reward_std": 0.4649389386177063, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3751562237739563, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 4174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 591.9375, "completions/mean_terminated_length": 591.9375, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.835, "grad_norm": 0.2084398716688156, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49385179.0, "reward": 0.9875165820121765, "reward_std": 0.47807252407073975, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5375312566757202, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 813.4375, "completions/mean_terminated_length": 783.357177734375, "completions/min_length": 464.0, "completions/min_terminated_length": 464.0, "epoch": 0.8352, "grad_norm": 0.1508539468050003, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49403714.0, "reward": 0.7807034254074097, "reward_std": 0.3338908553123474, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6864062547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 4176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 277.1875, "completions/mean_terminated_length": 277.1875, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.8354, "grad_norm": 0.0007010766421444714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49411957.0, "reward": 1.4995572566986084, "reward_std": 0.0004833567072637379, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991124868392944, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737500548362732, "step": 4177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 607.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 459.375, "completions/mean_terminated_length": 459.375, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.8356, "grad_norm": 0.22233402729034424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49423259.0, "reward": 1.139894962310791, "reward_std": 0.4795529544353485, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6547874808311462, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 4178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 764.0625, "completions/mean_terminated_length": 746.7333984375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.8358, "grad_norm": 0.33342283964157104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49439804.0, "reward": 0.9835785031318665, "reward_std": 0.3632873296737671, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.8421562910079956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3668749928474426, "step": 4179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 522.5625, "completions/mean_terminated_length": 489.13336181640625, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.836, "grad_norm": 0.049345266073942184, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49452229.0, "reward": 0.51734459400177, "reward_std": 0.1384401023387909, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 229.0, "completions/mean_terminated_length": 229.0, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.8362, "grad_norm": 0.23426011204719543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49459477.0, "reward": 1.4372386932373047, "reward_std": 0.249930277466774, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9369750022888184, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "step": 4181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 515.0, "completions/max_terminated_length": 515.0, "completions/mean_length": 331.125, "completions/mean_terminated_length": 331.125, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.8364, "grad_norm": 0.243342325091362, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49468487.0, "reward": 1.308072805404663, "reward_std": 0.357952356338501, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8661437630653381, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8818750381469727, "step": 4182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 537.25, "completions/mean_terminated_length": 537.25, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.8366, "grad_norm": 0.11108613014221191, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49481211.0, "reward": 0.8625004887580872, "reward_std": 0.18218214809894562, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6625000238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4937499761581421, "step": 4183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 639.1875, "completions/mean_terminated_length": 613.5333862304688, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.8368, "grad_norm": 0.2672032117843628, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49495310.0, "reward": 0.6717977523803711, "reward_std": 0.3732107877731323, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2810937464237213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 4184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 721.0625, "completions/mean_terminated_length": 677.7857666015625, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.837, "grad_norm": 0.3159016966819763, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49511263.0, "reward": 1.0767165422439575, "reward_std": 0.5869117379188538, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6534312963485718, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8306249380111694, "step": 4185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 593.125, "completions/mean_terminated_length": 593.125, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.8372, "grad_norm": 0.2159544825553894, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49524209.0, "reward": 1.142500877380371, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6600000262260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 381.0, "completions/max_terminated_length": 381.0, "completions/mean_length": 263.75, "completions/mean_terminated_length": 263.75, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.8374, "grad_norm": 0.575005054473877, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49532589.0, "reward": 1.0148072242736816, "reward_std": 0.5002864003181458, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5296125411987305, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 568.25, "completions/mean_terminated_length": 568.25, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.8376, "grad_norm": 0.2908514142036438, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49545153.0, "reward": 0.9492947459220886, "reward_std": 0.4572944939136505, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5235875248908997, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8650000095367432, "step": 4188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 321.0, "completions/mean_length": 410.875, "completions/mean_terminated_length": 269.3846130371094, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.8378, "grad_norm": 0.44458046555519104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49555775.0, "reward": 1.217941403388977, "reward_std": 0.6042683124542236, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8108812570571899, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.778124988079071, "step": 4189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 298.4375, "completions/mean_terminated_length": 298.4375, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.838, "grad_norm": 0.17266547679901123, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49564102.0, "reward": 0.6272447109222412, "reward_std": 0.25985902547836304, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19198749959468842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9000000357627869, "step": 4190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 308.3125, "completions/mean_terminated_length": 308.3125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.8382, "grad_norm": 0.38802069425582886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49572891.0, "reward": 1.439741611480713, "reward_std": 0.23759812116622925, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419812560081482, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 383.5625, "completions/mean_terminated_length": 383.5625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.8384, "grad_norm": 0.43647071719169617, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49582708.0, "reward": 1.0807008743286133, "reward_std": 0.4896578788757324, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5989000201225281, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 774.8125, "completions/mean_terminated_length": 739.2142944335938, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.8386, "grad_norm": 0.17578987777233124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49600321.0, "reward": 0.8426567912101746, "reward_std": 0.4543529152870178, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6228125095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.48124998807907104, "step": 4193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 562.8125, "completions/mean_terminated_length": 456.3846435546875, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.8388, "grad_norm": 0.6342144012451172, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49614158.0, "reward": 0.8464069366455078, "reward_std": 0.584985077381134, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5053125023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.699999988079071, "step": 4194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 981.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 566.625, "completions/mean_terminated_length": 566.625, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.839, "grad_norm": 0.21863801777362823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49628456.0, "reward": 0.8345320224761963, "reward_std": 0.38996031880378723, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4190624952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "step": 4195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 812.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 508.125, "completions/mean_terminated_length": 508.125, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.8392, "grad_norm": 0.3084668517112732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49641514.0, "reward": 1.3485593795776367, "reward_std": 0.4244302213191986, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8846167325973511, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8971250057220459, "step": 4196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 387.0625, "completions/mean_terminated_length": 387.0625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.8394, "grad_norm": 0.3090404272079468, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49651627.0, "reward": 1.1998447179794312, "reward_std": 0.44747036695480347, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7121875286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 4197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 418.0625, "completions/mean_terminated_length": 418.0625, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.8396, "grad_norm": 0.24858318269252777, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49662556.0, "reward": 1.3802103996276855, "reward_std": 0.32456883788108826, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854187726974487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 4198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 740.75, "completions/mean_terminated_length": 675.3846435546875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.8398, "grad_norm": 0.46592047810554504, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49678440.0, "reward": 0.9382039308547974, "reward_std": 0.6221157908439636, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5639062523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.746874988079071, "step": 4199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 759.6875, "completions/mean_terminated_length": 742.0667114257812, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.84, "grad_norm": 0.41450682282447815, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49695203.0, "reward": 0.8374757766723633, "reward_std": 0.49543294310569763, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4249500036239624, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8412500023841858, "step": 4200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 271.0, "completions/max_terminated_length": 271.0, "completions/mean_length": 208.4375, "completions/mean_terminated_length": 208.4375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.8402, "grad_norm": 0.32028427720069885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49702314.0, "reward": 1.3802071809768677, "reward_std": 0.324567586183548, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854125142097473, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 796.9375, "completions/mean_terminated_length": 693.727294921875, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.8404, "grad_norm": 0.2762640714645386, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49719433.0, "reward": 0.7161720991134644, "reward_std": 0.5138106346130371, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6823437809944153, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.109375, "step": 4202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1013.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 599.375, "completions/mean_terminated_length": 599.375, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.8406, "grad_norm": 0.0968267172574997, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49733375.0, "reward": 0.6459382772445679, "reward_std": 0.17838682234287262, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2918750047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 4203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 899.3125, "completions/mean_terminated_length": 870.5385131835938, "completions/min_length": 780.0, "completions/min_terminated_length": 780.0, "epoch": 0.8408, "grad_norm": 0.524588406085968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49751524.0, "reward": 0.7409257292747498, "reward_std": 0.567205011844635, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3568499982357025, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7737500071525574, "step": 4204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 267.125, "completions/mean_terminated_length": 267.125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.841, "grad_norm": 0.6190459132194519, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49760118.0, "reward": 0.9025759696960449, "reward_std": 0.47703468799591064, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43015000224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 354.6875, "completions/mean_terminated_length": 354.6875, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.8412, "grad_norm": 0.16222454607486725, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49769985.0, "reward": 1.4401042461395264, "reward_std": 0.2376948595046997, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9427062273025513, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9743750095367432, "step": 4206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 617.4375, "completions/mean_terminated_length": 590.3333740234375, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.8414, "grad_norm": 0.12784381210803986, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49784344.0, "reward": 0.8119536638259888, "reward_std": 0.32995644211769104, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6239062547683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4593749940395355, "step": 4207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 611.375, "completions/mean_terminated_length": 583.86669921875, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.8416, "grad_norm": 0.22895237803459167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49798078.0, "reward": 0.6923446655273438, "reward_std": 0.4220276176929474, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25968748331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 690.5625, "completions/mean_terminated_length": 613.6154174804688, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.8418, "grad_norm": 0.5740893483161926, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49812647.0, "reward": 0.7631255984306335, "reward_std": 0.5080060958862305, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5262500047683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5062500238418579, "step": 4209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 436.375, "completions/mean_terminated_length": 397.20001220703125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.842, "grad_norm": 0.24063900113105774, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49823805.0, "reward": 0.9456258416175842, "reward_std": 0.5251056551933289, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5162500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 4210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 364.9375, "completions/mean_terminated_length": 364.9375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.8422, "grad_norm": 0.18764887750148773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49833132.0, "reward": 1.341147780418396, "reward_std": 0.34608736634254456, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8697937726974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256249666213989, "step": 4211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 617.6875, "completions/mean_terminated_length": 617.6875, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.8424, "grad_norm": 0.4164898097515106, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49847447.0, "reward": 1.1125789880752563, "reward_std": 0.4209938645362854, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7251562476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.809374988079071, "step": 4212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 303.5, "completions/mean_terminated_length": 303.5, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.8426, "grad_norm": 0.3982533812522888, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49856415.0, "reward": 0.7814759016036987, "reward_std": 0.42807453870773315, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3129499852657318, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9575000405311584, "step": 4213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 270.125, "completions/mean_terminated_length": 270.125, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.8428, "grad_norm": 0.21969850361347198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49864385.0, "reward": 0.6332696676254272, "reward_std": 0.2549535036087036, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20403750240802765, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 4214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 486.875, "completions/mean_terminated_length": 410.14288330078125, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.843, "grad_norm": 0.242422416806221, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49875743.0, "reward": 1.192735195159912, "reward_std": 0.5656371712684631, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7604687809944153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 4215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 605.75, "completions/mean_terminated_length": 605.75, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.8432, "grad_norm": 0.15592339634895325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49889803.0, "reward": 1.259441614151001, "reward_std": 0.4281385540962219, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7688812613487244, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 435.1875, "completions/mean_terminated_length": 435.1875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.8434, "grad_norm": 0.23632550239562988, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49900622.0, "reward": 1.0216259956359863, "reward_std": 0.49279817938804626, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5432499647140503, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "step": 4217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 321.125, "completions/mean_terminated_length": 321.125, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.8436, "grad_norm": 0.16547229886054993, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49909664.0, "reward": 1.3799633979797363, "reward_std": 0.32447391748428345, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849250078201294, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 360.0625, "completions/mean_terminated_length": 360.0625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.8438, "grad_norm": 0.26056969165802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49918945.0, "reward": 0.7863259315490723, "reward_std": 0.42498716711997986, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3226499855518341, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 452.0, "completions/max_terminated_length": 452.0, "completions/mean_length": 305.875, "completions/mean_terminated_length": 305.875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.844, "grad_norm": 0.41509461402893066, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49928143.0, "reward": 0.8381853699684143, "reward_std": 0.4598574638366699, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36386874318122864, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 4220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 708.5625, "completions/mean_terminated_length": 635.7692260742188, "completions/min_length": 513.0, "completions/min_terminated_length": 513.0, "epoch": 0.8442, "grad_norm": 0.13715001940727234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49944472.0, "reward": 0.9209257960319519, "reward_std": 0.6276492476463318, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5293499827384949, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7737500071525574, "step": 4221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 269.375, "completions/mean_terminated_length": 269.375, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.8444, "grad_norm": 0.00036151695530861616, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49955118.0, "reward": 1.499119758605957, "reward_std": 0.00056636449880898, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982374906539917, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 4222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 388.9375, "completions/mean_terminated_length": 388.9375, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.8446, "grad_norm": 0.5332556366920471, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49965117.0, "reward": 1.043725848197937, "reward_std": 0.4746720790863037, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5874500274658203, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287499785423279, "step": 4223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 442.8125, "completions/mean_terminated_length": 442.8125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.8448, "grad_norm": 0.11604531854391098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49976458.0, "reward": 1.3936572074890137, "reward_std": 0.28849631547927856, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9123125076293945, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9387500286102295, "step": 4224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 443.0, "completions/mean_terminated_length": 443.0, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.845, "grad_norm": 0.1986137330532074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49987578.0, "reward": 1.3802196979522705, "reward_std": 0.32457253336906433, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8854374885559082, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 266.125, "completions/mean_terminated_length": 266.125, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.8452, "grad_norm": 0.2944301664829254, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49995948.0, "reward": 0.7119573354721069, "reward_std": 0.39129000902175903, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23641249537467957, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 4226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 736.4375, "completions/mean_terminated_length": 736.4375, "completions/min_length": 516.0, "completions/min_terminated_length": 516.0, "epoch": 0.8454, "grad_norm": 0.18817152082920074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50011747.0, "reward": 1.374350905418396, "reward_std": 0.28276169300079346, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9362000226974487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8537499904632568, "step": 4227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 640.5, "completions/mean_terminated_length": 640.5, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.8456, "grad_norm": 0.159254252910614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50025819.0, "reward": 1.1034257411956787, "reward_std": 0.3794747292995453, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.769349992275238, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7425000071525574, "step": 4228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 581.8125, "completions/mean_terminated_length": 552.3333740234375, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.8458, "grad_norm": 0.14855709671974182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50038952.0, "reward": 0.9923446178436279, "reward_std": 0.540053129196167, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5471875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 371.0625, "completions/mean_terminated_length": 371.0625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.846, "grad_norm": 0.4560684561729431, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50049545.0, "reward": 1.0833353996276855, "reward_std": 0.486921101808548, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6041687726974487, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "step": 4230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 682.0625, "completions/mean_terminated_length": 659.2667236328125, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.8462, "grad_norm": 0.28198301792144775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50064362.0, "reward": 0.9073448181152344, "reward_std": 0.5077172517776489, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5021874904632568, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 4231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 504.8125, "completions/mean_terminated_length": 470.20001220703125, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.8464, "grad_norm": 0.228195920586586, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50075959.0, "reward": 0.7021100521087646, "reward_std": 0.2904317378997803, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46671873331069946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.546875, "step": 4232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 837.0, "completions/max_terminated_length": 837.0, "completions/mean_length": 561.4375, "completions/mean_terminated_length": 561.4375, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8466, "grad_norm": 0.10229763388633728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50088654.0, "reward": 0.7046882510185242, "reward_std": 0.2738114893436432, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.34687501192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7874999642372131, "step": 4233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 414.5625, "completions/mean_terminated_length": 414.5625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.8468, "grad_norm": 0.31174349784851074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50098759.0, "reward": 0.8421885967254639, "reward_std": 0.45713528990745544, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37187498807907104, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 795.375, "completions/mean_terminated_length": 780.1333618164062, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.847, "grad_norm": 0.08306087553501129, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50116125.0, "reward": 0.832109808921814, "reward_std": 0.23635156452655792, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7267187833786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.40312498807907104, "step": 4235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 507.0, "completions/mean_terminated_length": 472.5333557128906, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.8472, "grad_norm": 0.18815824389457703, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50127965.0, "reward": 0.8568755984306335, "reward_std": 0.39920753240585327, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5262500047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6312500238418579, "step": 4236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 583.0, "completions/mean_terminated_length": 553.6000366210938, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.8474, "grad_norm": 0.18834097683429718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50141229.0, "reward": 0.6826568841934204, "reward_std": 0.27116966247558594, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.42781251668930054, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6062500476837158, "step": 4237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 766.0, "completions/mean_length": 594.875, "completions/mean_terminated_length": 533.5714721679688, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.8476, "grad_norm": 0.19734779000282288, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50154299.0, "reward": 0.6314852833747864, "reward_std": 0.3670397996902466, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3254687488079071, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6468749642372131, "step": 4238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 431.0, "completions/mean_terminated_length": 391.4666748046875, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.8478, "grad_norm": 0.045424848794937134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50165323.0, "reward": 0.546629011631012, "reward_std": 0.18435443937778473, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15575623512268066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8268750309944153, "step": 4239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 673.125, "completions/mean_terminated_length": 673.125, "completions/min_length": 510.0, "completions/min_terminated_length": 510.0, "epoch": 0.848, "grad_norm": 0.21259300410747528, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50180029.0, "reward": 1.0853915214538574, "reward_std": 0.44812217354774475, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6707812547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "step": 4240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 670.4375, "completions/mean_terminated_length": 588.84619140625, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.8482, "grad_norm": 0.2904350757598877, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50195140.0, "reward": 0.8017851114273071, "reward_std": 0.5655889511108398, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.47856876254081726, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6581249833106995, "step": 4241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 476.25, "completions/mean_terminated_length": 439.7333679199219, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.8484, "grad_norm": 0.22311484813690186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50206760.0, "reward": 1.3130476474761963, "reward_std": 0.43711936473846436, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8760937452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 4242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 244.375, "completions/mean_terminated_length": 244.375, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.8486, "grad_norm": 0.23609484732151031, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50214734.0, "reward": 0.875625729560852, "reward_std": 0.19769352674484253, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7512500286102295, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3374999761581421, "step": 4243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 494.375, "completions/mean_terminated_length": 494.375, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.8488, "grad_norm": 0.3867432475090027, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50226420.0, "reward": 1.1422665119171143, "reward_std": 0.4748133718967438, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6595312356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 673.4375, "completions/mean_terminated_length": 673.4375, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.849, "grad_norm": 0.19738392531871796, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50241099.0, "reward": 0.7364071607589722, "reward_std": 0.3312157988548279, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3478125035762787, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 4245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 413.0, "completions/mean_terminated_length": 413.0, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.8492, "grad_norm": 0.1458187699317932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50252075.0, "reward": 0.6675009727478027, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20999999344348907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 374.75, "completions/mean_terminated_length": 374.75, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.8494, "grad_norm": 0.18822643160820007, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50262119.0, "reward": 0.8737474679946899, "reward_std": 0.2085118293762207, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7474937438964844, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3368750214576721, "step": 4247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 530.75, "completions/mean_terminated_length": 530.75, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.8496, "grad_norm": 0.16203011572360992, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50275555.0, "reward": 1.0844290256500244, "reward_std": 0.4761519134044647, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6063562631607056, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9193750023841858, "step": 4248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 409.5625, "completions/mean_terminated_length": 409.5625, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.8498, "grad_norm": 0.17996647953987122, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50285788.0, "reward": 1.2891414165496826, "reward_std": 0.3880109190940857, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8282812833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 852.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 448.4375, "completions/mean_terminated_length": 448.4375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.85, "grad_norm": 0.12896184623241425, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50298755.0, "reward": 1.4378509521484375, "reward_std": 0.24481385946273804, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9381999969482422, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 300.75, "completions/mean_terminated_length": 300.75, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.8502, "grad_norm": 0.12507528066635132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50307919.0, "reward": 1.439866542816162, "reward_std": 0.23763154447078705, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.942231297492981, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "step": 4251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 407.875, "completions/mean_terminated_length": 407.875, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.8504, "grad_norm": 0.25204822421073914, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50318333.0, "reward": 0.8172667026519775, "reward_std": 0.4159919023513794, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.38453125953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 470.1875, "completions/mean_terminated_length": 470.1875, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.8506, "grad_norm": 0.35054922103881836, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50329248.0, "reward": 0.7960821390151978, "reward_std": 0.3768145740032196, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.40466248989105225, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8424999713897705, "step": 4253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 270.0, "completions/mean_terminated_length": 270.0, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.8508, "grad_norm": 0.511290967464447, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50336864.0, "reward": 0.9548853635787964, "reward_std": 0.49606192111968994, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4722687602043152, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 4254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 653.5625, "completions/mean_terminated_length": 628.86669921875, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.851, "grad_norm": 0.2693762481212616, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50351113.0, "reward": 1.068347692489624, "reward_std": 0.4968685507774353, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6991937160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7518750429153442, "step": 4255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 534.875, "completions/mean_terminated_length": 534.875, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.8512, "grad_norm": 0.24522122740745544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50363751.0, "reward": 1.0606133937835693, "reward_std": 0.4596783220767975, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6212249994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9049999713897705, "step": 4256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 360.0, "completions/max_terminated_length": 360.0, "completions/mean_length": 282.3125, "completions/mean_terminated_length": 282.3125, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.8514, "grad_norm": 0.29643070697784424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50372572.0, "reward": 0.7724291086196899, "reward_std": 0.4335106611251831, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2948562502861023, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 4257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 720.75, "completions/mean_terminated_length": 720.75, "completions/min_length": 478.0, "completions/min_terminated_length": 478.0, "epoch": 0.8516, "grad_norm": 0.08120401203632355, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50388648.0, "reward": 0.6783602237701416, "reward_std": 0.2517176568508148, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2942187190055847, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8468749523162842, "step": 4258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 425.0, "completions/max_terminated_length": 425.0, "completions/mean_length": 310.25, "completions/mean_terminated_length": 310.25, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.8518, "grad_norm": 0.0008110472117550671, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50397244.0, "reward": 1.499463438987732, "reward_std": 0.0005731788114644587, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998924970626831, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 4259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 629.9375, "completions/mean_terminated_length": 629.9375, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.852, "grad_norm": 0.18422339856624603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50412187.0, "reward": 0.7706258893013, "reward_std": 0.33003658056259155, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.41624999046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 4260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 435.75, "completions/mean_terminated_length": 396.5333557128906, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.8522, "grad_norm": 0.47636494040489197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50423351.0, "reward": 1.0006102323532104, "reward_std": 0.5354099273681641, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5637187957763672, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8793749809265137, "step": 4261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 552.0, "completions/max_terminated_length": 552.0, "completions/mean_length": 379.4375, "completions/mean_terminated_length": 379.4375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.8524, "grad_norm": 0.6506142020225525, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50433406.0, "reward": 1.081626057624817, "reward_std": 0.48918023705482483, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6007499694824219, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 477.875, "completions/mean_terminated_length": 477.875, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.8526, "grad_norm": 0.17879147827625275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50445084.0, "reward": 1.3734383583068848, "reward_std": 0.28292587399482727, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9343750476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 4263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 669.125, "completions/mean_terminated_length": 645.4666748046875, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.8528, "grad_norm": 0.14057078957557678, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50462510.0, "reward": 0.8974223136901855, "reward_std": 0.36195090413093567, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.7323437333106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3968750238418579, "step": 4264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 501.5, "completions/mean_terminated_length": 501.5, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.853, "grad_norm": 0.3602672219276428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50474518.0, "reward": 1.0515539646148682, "reward_std": 0.4014333486557007, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7281062602996826, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6731250286102295, "step": 4265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 545.375, "completions/mean_terminated_length": 545.375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.8532, "grad_norm": 0.41804784536361694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50487564.0, "reward": 0.9108477830886841, "reward_std": 0.47070497274398804, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4466937780380249, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 339.0625, "completions/mean_terminated_length": 339.0625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.8534, "grad_norm": 0.1939162015914917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50497405.0, "reward": 0.6092071533203125, "reward_std": 0.237626850605011, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15591248869895935, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 621.0625, "completions/mean_terminated_length": 621.0625, "completions/min_length": 449.0, "completions/min_terminated_length": 449.0, "epoch": 0.8536, "grad_norm": 0.21930785477161407, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50511710.0, "reward": 0.8134382963180542, "reward_std": 0.37413936853408813, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.43937501311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8125, "step": 4268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 446.375, "completions/mean_terminated_length": 446.375, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.8538, "grad_norm": 0.20563513040542603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50522308.0, "reward": 0.9583978652954102, "reward_std": 0.49245700240135193, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47929373383522034, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1002.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 787.1875, "completions/mean_terminated_length": 787.1875, "completions/min_length": 571.0, "completions/min_terminated_length": 571.0, "epoch": 0.854, "grad_norm": 0.1230730265378952, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50539191.0, "reward": 0.7489068508148193, "reward_std": 0.18391817808151245, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49781250953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.637499988079071, "step": 4270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 288.875, "completions/mean_terminated_length": 288.875, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.8542, "grad_norm": 0.00028480819310061634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50547589.0, "reward": 1.4996322393417358, "reward_std": 0.0005278627504594624, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9992625117301941, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "step": 4271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 566.0, "completions/mean_terminated_length": 566.0, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.8544, "grad_norm": 0.07859624922275543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50561605.0, "reward": 0.9406255483627319, "reward_std": 0.1544543206691742, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8812500238418579, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.17500001192092896, "step": 4272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 262.375, "completions/mean_terminated_length": 262.375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.8546, "grad_norm": 0.2777708172798157, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50569867.0, "reward": 0.9274224638938904, "reward_std": 0.2855696976184845, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.7298437356948853, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.43437498807907104, "step": 4273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 812.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 504.75, "completions/mean_terminated_length": 504.75, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.8548, "grad_norm": 0.30042198300361633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50582135.0, "reward": 1.2081257104873657, "reward_std": 0.3499126136302948, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.8537499904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.75, "step": 4274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1004.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 703.875, "completions/mean_terminated_length": 703.875, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.855, "grad_norm": 0.2126668393611908, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50597285.0, "reward": 1.29007887840271, "reward_std": 0.32865750789642334, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8301562070846558, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 4275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 969.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 666.0, "completions/mean_terminated_length": 666.0, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.8552, "grad_norm": 0.22081954777240753, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50611909.0, "reward": 0.7634382843971252, "reward_std": 0.3227793574333191, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.40187501907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7999999523162842, "step": 4276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 620.875, "completions/mean_terminated_length": 527.84619140625, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.8554, "grad_norm": 0.5515233278274536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50625907.0, "reward": 0.8312506675720215, "reward_std": 0.5899165272712708, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4749999940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 4277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 436.375, "completions/mean_terminated_length": 352.4285888671875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.8556, "grad_norm": 1.3351233005523682, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50636825.0, "reward": 0.8303757905960083, "reward_std": 0.5213509798049927, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.47325000166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7699999809265137, "step": 4278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 276.8125, "completions/mean_terminated_length": 276.8125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.8558, "grad_norm": 0.6337725520133972, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50644774.0, "reward": 1.213204026222229, "reward_std": 0.4370337724685669, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7389062643051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 673.3125, "completions/mean_terminated_length": 673.3125, "completions/min_length": 511.0, "completions/min_terminated_length": 511.0, "epoch": 0.856, "grad_norm": 0.18655410408973694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50659387.0, "reward": 0.798269510269165, "reward_std": 0.3027712106704712, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4715374708175659, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7800000309944153, "step": 4280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 622.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 446.5625, "completions/mean_terminated_length": 446.5625, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.8562, "grad_norm": 0.07545077800750732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50670868.0, "reward": 0.6055448055267334, "reward_std": 0.23901356756687164, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14858749508857727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 492.4375, "completions/mean_terminated_length": 492.4375, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.8564, "grad_norm": 0.034541111439466476, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50682987.0, "reward": 0.9434378147125244, "reward_std": 0.07354618608951569, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8868750333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.26875001192092896, "step": 4282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 406.875, "completions/mean_terminated_length": 406.875, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.8566, "grad_norm": 0.4306965470314026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50693417.0, "reward": 1.145391583442688, "reward_std": 0.47127190232276917, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6657812595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 963.5, "completions/mean_terminated_length": 830.4000244140625, "completions/min_length": 713.0, "completions/min_terminated_length": 713.0, "epoch": 0.8568, "grad_norm": 0.5086621046066284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50713569.0, "reward": 0.21726584434509277, "reward_std": 0.348880410194397, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12203124910593033, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.22812500596046448, "step": 4284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 342.125, "completions/mean_terminated_length": 342.125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.857, "grad_norm": 0.31831642985343933, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50722963.0, "reward": 0.7564852237701416, "reward_std": 0.3667188286781311, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3254687488079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 542.25, "completions/mean_terminated_length": 510.13336181640625, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.8572, "grad_norm": 0.3439129889011383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50735479.0, "reward": 0.8760164976119995, "reward_std": 0.5159159898757935, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43953126668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.878125011920929, "step": 4286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 410.0, "completions/mean_length": 449.4375, "completions/mean_terminated_length": 367.3571472167969, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.8574, "grad_norm": 0.4539811909198761, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50746430.0, "reward": 0.9454693794250488, "reward_std": 0.522553563117981, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5159374475479126, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.612500011920929, "step": 4287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 363.5, "completions/mean_terminated_length": 363.5, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.8576, "grad_norm": 0.0003201069193892181, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50756070.0, "reward": 1.4995572566986084, "reward_std": 0.0005635914858430624, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9991124868392944, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762500524520874, "step": 4288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 690.5625, "completions/mean_terminated_length": 579.4166870117188, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.8578, "grad_norm": 0.23112745583057404, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50771551.0, "reward": 0.832097589969635, "reward_std": 0.5882300734519958, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6016937494277954, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.43937498331069946, "step": 4289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 321.6875, "completions/mean_terminated_length": 321.6875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.858, "grad_norm": 0.2558160126209259, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50780458.0, "reward": 0.6008884310722351, "reward_std": 0.2397720217704773, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13927499949932098, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 348.0, "completions/mean_terminated_length": 348.0, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.8582, "grad_norm": 0.2153916209936142, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50790410.0, "reward": 1.4396477937698364, "reward_std": 0.23757296800613403, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9417937397956848, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 732.0, "completions/max_terminated_length": 732.0, "completions/mean_length": 605.0, "completions/mean_terminated_length": 605.0, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.8584, "grad_norm": 0.014130310155451298, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50804970.0, "reward": 0.5742977857589722, "reward_std": 0.03676489368081093, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1485937535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 4292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1008.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 506.9375, "completions/mean_terminated_length": 506.9375, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.8586, "grad_norm": 0.22601386904716492, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50816745.0, "reward": 0.812729001045227, "reward_std": 0.4166997969150543, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3754562735557556, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143750071525574, "step": 4293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 552.4375, "completions/mean_terminated_length": 443.6153869628906, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.8588, "grad_norm": 0.6457797288894653, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50829760.0, "reward": 1.0987508296966553, "reward_std": 0.6315998435020447, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6974999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.768750011920929, "step": 4294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 976.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 560.875, "completions/mean_terminated_length": 560.875, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.859, "grad_norm": 0.0533563531935215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50842750.0, "reward": 0.5947664976119995, "reward_std": 0.1008475199341774, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18953125178813934, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 534.1875, "completions/mean_terminated_length": 501.5333557128906, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.8592, "grad_norm": 0.19104772806167603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50855073.0, "reward": 0.5746102333068848, "reward_std": 0.2007163017988205, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21171873807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 4296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 410.0, "completions/mean_terminated_length": 410.0, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8594, "grad_norm": 0.14020441472530365, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50865537.0, "reward": 1.4004571437835693, "reward_std": 0.27531546354293823, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9259124994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237500429153442, "step": 4297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 561.75, "completions/mean_terminated_length": 530.933349609375, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.8596, "grad_norm": 0.13746050000190735, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50878573.0, "reward": 0.7942162752151489, "reward_std": 0.38173097372055054, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4634312391281128, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6368750333786011, "step": 4298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 640.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 506.9375, "completions/mean_terminated_length": 506.9375, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.8598, "grad_norm": 0.1432134360074997, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50890412.0, "reward": 1.2567977905273438, "reward_std": 0.4309200942516327, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7635937929153442, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 4299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 564.4375, "completions/mean_terminated_length": 533.800048828125, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.86, "grad_norm": 0.24776501953601837, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50903299.0, "reward": 1.0105102062225342, "reward_std": 0.5043582916259766, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5835187435150146, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7618750333786011, "step": 4300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 881.25, "completions/mean_terminated_length": 816.3636474609375, "completions/min_length": 676.0, "completions/min_terminated_length": 676.0, "epoch": 0.8602, "grad_norm": 0.18055179715156555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50922791.0, "reward": 0.5196099281311035, "reward_std": 0.4367577135562897, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2892187535762787, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5093749761581421, "step": 4301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 676.0, "completions/max_terminated_length": 676.0, "completions/mean_length": 517.125, "completions/mean_terminated_length": 517.125, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.8604, "grad_norm": 0.1431194692850113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50934889.0, "reward": 1.2186603546142578, "reward_std": 0.395062655210495, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8123186826705933, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8268749713897705, "step": 4302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 372.0, "completions/max_terminated_length": 372.0, "completions/mean_length": 302.9375, "completions/mean_terminated_length": 302.9375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.8606, "grad_norm": 0.48217713832855225, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50943592.0, "reward": 1.022472858428955, "reward_std": 0.49233323335647583, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5449437499046326, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 552.5, "completions/mean_terminated_length": 552.5, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.8608, "grad_norm": 0.036758750677108765, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50956416.0, "reward": 0.9493752717971802, "reward_std": 0.07082530111074448, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8987500071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.25, "step": 4304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 312.0, "completions/max_terminated_length": 312.0, "completions/mean_length": 242.25, "completions/mean_terminated_length": 242.25, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.861, "grad_norm": 0.2702966332435608, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50964036.0, "reward": 1.0831258296966553, "reward_std": 0.48673015832901, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6037499904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 296.625, "completions/mean_terminated_length": 296.625, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.8612, "grad_norm": 0.260951429605484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50972318.0, "reward": 1.3163197040557861, "reward_std": 0.38041162490844727, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8201375007629395, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9325000047683716, "step": 4306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 465.0, "completions/max_terminated_length": 465.0, "completions/mean_length": 359.625, "completions/mean_terminated_length": 359.625, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.8614, "grad_norm": 0.5610560774803162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50982184.0, "reward": 1.3208322525024414, "reward_std": 0.3830597996711731, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8291624784469604, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 308.125, "completions/mean_terminated_length": 308.125, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.8616, "grad_norm": 0.36805981397628784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50991370.0, "reward": 0.7844415903091431, "reward_std": 0.4259919226169586, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31888124346733093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 504.0, "completions/max_terminated_length": 504.0, "completions/mean_length": 287.1875, "completions/mean_terminated_length": 287.1875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.8618, "grad_norm": 0.32458293437957764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50999965.0, "reward": 1.1371634006500244, "reward_std": 0.40097445249557495, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7118250131607056, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7587499618530273, "step": 4309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 431.25, "completions/mean_terminated_length": 431.25, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.862, "grad_norm": 0.14610998332500458, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51011569.0, "reward": 0.6657571196556091, "reward_std": 0.26050522923469543, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2690124809741974, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 4310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 277.0, "completions/mean_terminated_length": 277.0, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.8622, "grad_norm": 0.2320854663848877, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51021809.0, "reward": 0.6482415795326233, "reward_std": 0.33249935507774353, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17148125171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 4311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 931.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 670.4375, "completions/mean_terminated_length": 670.4375, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.8624, "grad_norm": 0.08772680908441544, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51037320.0, "reward": 0.7182039618492126, "reward_std": 0.310369610786438, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3114062547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.796875, "step": 4312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 671.0, "completions/max_terminated_length": 671.0, "completions/mean_length": 531.625, "completions/mean_terminated_length": 531.625, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.8626, "grad_norm": 0.09371792525053024, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51050066.0, "reward": 0.6115633249282837, "reward_std": 0.14579173922538757, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22312499582767487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 4313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 506.4375, "completions/mean_terminated_length": 506.4375, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.8628, "grad_norm": 0.10913418233394623, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51061929.0, "reward": 0.6982821226119995, "reward_std": 0.3315140902996063, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27156251668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 4314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 751.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 475.6875, "completions/mean_terminated_length": 475.6875, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.863, "grad_norm": 0.18756015598773956, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51073348.0, "reward": 0.7912445068359375, "reward_std": 0.33051395416259766, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3949875235557556, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7549999952316284, "step": 4315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 973.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 598.9375, "completions/mean_terminated_length": 598.9375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.8632, "grad_norm": 0.05938129499554634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51087939.0, "reward": 0.5855478048324585, "reward_std": 0.08867646008729935, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17109373211860657, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 434.625, "completions/mean_terminated_length": 434.625, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.8634, "grad_norm": 0.27863675355911255, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51099501.0, "reward": 0.8821859359741211, "reward_std": 0.4941534399986267, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3893687427043915, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9856187105178833, "step": 4317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 561.4375, "completions/mean_terminated_length": 561.4375, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.8636, "grad_norm": 0.2216392159461975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51112420.0, "reward": 1.0797134637832642, "reward_std": 0.4912372827529907, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5969250202178955, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 4318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 453.125, "completions/mean_terminated_length": 453.125, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8638, "grad_norm": 0.24570584297180176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51123510.0, "reward": 1.0802323818206787, "reward_std": 0.4905763268470764, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.597962498664856, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 524.0625, "completions/mean_terminated_length": 524.0625, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.864, "grad_norm": 0.1277022510766983, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51135767.0, "reward": 1.3416414260864258, "reward_std": 0.32249027490615845, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8707813024520874, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 4320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 706.8125, "completions/mean_terminated_length": 685.6666870117188, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 0.8642, "grad_norm": 0.15863730013370514, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51151284.0, "reward": 0.6667194962501526, "reward_std": 0.32366839051246643, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3334375023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.731249988079071, "step": 4321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 548.375, "completions/mean_terminated_length": 516.6666870117188, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.8644, "grad_norm": 0.22440098226070404, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51164042.0, "reward": 0.9284102320671082, "reward_std": 0.5365000367164612, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4818187654018402, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 4322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 648.0, "completions/mean_terminated_length": 594.2857666015625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.8646, "grad_norm": 0.27225616574287415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51178074.0, "reward": 0.6637508869171143, "reward_std": 0.45539507269859314, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26499998569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.824999988079071, "step": 4323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 486.0, "completions/max_terminated_length": 486.0, "completions/mean_length": 386.0625, "completions/mean_terminated_length": 386.0625, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.8648, "grad_norm": 0.11641263216733932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51188411.0, "reward": 1.4316884279251099, "reward_std": 0.23746931552886963, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9258750081062317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9275000095367432, "step": 4324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 314.0, "completions/max_terminated_length": 314.0, "completions/mean_length": 255.9375, "completions/mean_terminated_length": 255.9375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.865, "grad_norm": 0.5682026147842407, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51195962.0, "reward": 0.725663423538208, "reward_std": 0.3842920958995819, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.263824999332428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 4325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 826.0, "completions/mean_length": 633.125, "completions/mean_terminated_length": 607.0667114257812, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.8652, "grad_norm": 0.1354752779006958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51210316.0, "reward": 0.6375792026519775, "reward_std": 0.3667922616004944, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21265624463558197, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 4326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 379.6875, "completions/mean_terminated_length": 379.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.8654, "grad_norm": 0.6039342284202576, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51220583.0, "reward": 0.9329696297645569, "reward_std": 0.4657990634441376, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4909375011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 4327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 777.875, "completions/mean_terminated_length": 695.8333740234375, "completions/min_length": 601.0, "completions/min_terminated_length": 601.0, "epoch": 0.8656, "grad_norm": 0.19378511607646942, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51238085.0, "reward": 0.6439065933227539, "reward_std": 0.4089812636375427, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5378124713897705, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.32499998807907104, "step": 4328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 532.9375, "completions/mean_terminated_length": 532.9375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.8658, "grad_norm": 0.12454324215650558, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51251636.0, "reward": 0.6871882677078247, "reward_std": 0.21217772364616394, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.37437498569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 4329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 677.0, "completions/max_terminated_length": 677.0, "completions/mean_length": 432.9375, "completions/mean_terminated_length": 432.9375, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.866, "grad_norm": 0.3446209132671356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51262355.0, "reward": 0.7366290092468262, "reward_std": 0.3787807524204254, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2857562303543091, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9393749833106995, "step": 4330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 352.0, "completions/max_terminated_length": 352.0, "completions/mean_length": 269.9375, "completions/mean_terminated_length": 269.9375, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.8662, "grad_norm": 0.12783880531787872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51270226.0, "reward": 1.440054178237915, "reward_std": 0.2376815676689148, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9426062703132629, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731250405311584, "step": 4331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 914.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 502.9375, "completions/mean_terminated_length": 502.9375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.8664, "grad_norm": 0.15455608069896698, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51282433.0, "reward": 0.7512508630752563, "reward_std": 0.2900817394256592, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4399999976158142, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6624999642372131, "step": 4332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 680.5625, "completions/mean_terminated_length": 657.6666870117188, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.8666, "grad_norm": 0.16754429042339325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51298218.0, "reward": 0.6044540405273438, "reward_std": 0.30108413100242615, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20890624821186066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 4333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 408.125, "completions/mean_terminated_length": 367.0666809082031, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.8668, "grad_norm": 0.31075319647789, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51308460.0, "reward": 1.1675760746002197, "reward_std": 0.5234941244125366, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7101500034332275, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 4334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 251.0, "completions/max_terminated_length": 251.0, "completions/mean_length": 167.25, "completions/mean_terminated_length": 167.25, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "epoch": 0.867, "grad_norm": 0.04100579395890236, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51314608.0, "reward": 0.526066780090332, "reward_std": 0.03173001483082771, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.052131183445453644, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9730499982833862, "step": 4335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 774.375, "completions/mean_terminated_length": 691.1666870117188, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.8672, "grad_norm": 0.20118628442287445, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51330662.0, "reward": 0.6820316314697266, "reward_std": 0.4442232847213745, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5515625476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3999999761581421, "step": 4336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 388.3125, "completions/mean_terminated_length": 388.3125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.8674, "grad_norm": 0.2867799401283264, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51340715.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 732.375, "completions/mean_terminated_length": 690.7142944335938, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.8676, "grad_norm": 0.32259875535964966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51357281.0, "reward": 0.7559888362884521, "reward_std": 0.5446584820747375, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3244749903678894, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.862500011920929, "step": 4338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 416.0, "completions/mean_terminated_length": 375.4666748046875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.8678, "grad_norm": 0.07648927718400955, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51367745.0, "reward": 0.5171884298324585, "reward_std": 0.13948550820350647, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09687499701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 391.125, "completions/mean_terminated_length": 391.125, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.868, "grad_norm": 0.12121202796697617, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51378451.0, "reward": 0.6095258593559265, "reward_std": 0.13930930197238922, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21904999017715454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8612499833106995, "step": 4340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 995.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 474.625, "completions/mean_terminated_length": 474.625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.8682, "grad_norm": 0.5917928814888, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51389997.0, "reward": 1.0586726665496826, "reward_std": 0.3872562646865845, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7423437237739563, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6656249761581421, "step": 4341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 562.625, "completions/mean_terminated_length": 496.71429443359375, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.8684, "grad_norm": 0.21155966818332672, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51402775.0, "reward": 0.5269477367401123, "reward_std": 0.24734140932559967, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17889374494552612, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7518749833106995, "step": 4342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 249.125, "completions/mean_terminated_length": 249.125, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.8686, "grad_norm": 0.0015615475131198764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51410569.0, "reward": 1.4988884925842285, "reward_std": 0.0011796074686571956, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977749586105347, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 4343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 476.0, "completions/mean_terminated_length": 476.0, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.8688, "grad_norm": 0.1808890998363495, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51425689.0, "reward": 0.6008884310722351, "reward_std": 0.2397719770669937, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13927499949932098, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 533.375, "completions/mean_terminated_length": 533.375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.869, "grad_norm": 0.14884087443351746, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51438143.0, "reward": 0.7862509489059448, "reward_std": 0.4248528778553009, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 669.5625, "completions/mean_terminated_length": 508.4545593261719, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.8692, "grad_norm": 0.3657638728618622, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51452264.0, "reward": 0.6747663021087646, "reward_std": 0.6119806170463562, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34953126311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.621874988079071, "step": 4346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 705.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 463.8125, "completions/mean_terminated_length": 463.8125, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.8694, "grad_norm": 0.11682330816984177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51463781.0, "reward": 1.3209540843963623, "reward_std": 0.3831203281879425, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8294062614440918, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 460.3125, "completions/mean_terminated_length": 460.3125, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.8696, "grad_norm": 0.14464552700519562, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51474922.0, "reward": 0.9587504863739014, "reward_std": 0.270589679479599, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7300000190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4125000238418579, "step": 4348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 789.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 544.875, "completions/mean_terminated_length": 544.875, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.8698, "grad_norm": 0.07380706071853638, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51487784.0, "reward": 0.7992974519729614, "reward_std": 0.14528535306453705, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5985937118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5843750238418579, "step": 4349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 417.3125, "completions/mean_terminated_length": 417.3125, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.87, "grad_norm": 0.3482055068016052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51498541.0, "reward": 0.863394558429718, "reward_std": 0.45935359597206116, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4767875075340271, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7699999809265137, "step": 4350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 358.0625, "completions/mean_terminated_length": 358.0625, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.8702, "grad_norm": 0.06847280263900757, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51508222.0, "reward": 0.5660135746002197, "reward_std": 0.07596854865550995, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13202500343322754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9275000095367432, "step": 4351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 393.1875, "completions/mean_terminated_length": 393.1875, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.8704, "grad_norm": 0.42074528336524963, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51518193.0, "reward": 1.0478541851043701, "reward_std": 0.47073814272880554, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5957062244415283, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 4352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 527.6875, "completions/mean_terminated_length": 527.6875, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.8706, "grad_norm": 0.175984725356102, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51530300.0, "reward": 1.0007038116455078, "reward_std": 0.3190215229988098, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.7514062523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5718749761581421, "step": 4353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 417.3125, "completions/mean_terminated_length": 417.3125, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.8708, "grad_norm": 0.3986028730869293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51540545.0, "reward": 0.8456915616989136, "reward_std": 0.45487847924232483, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3788812458515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 809.5625, "completions/mean_terminated_length": 712.0909423828125, "completions/min_length": 574.0, "completions/min_terminated_length": 574.0, "epoch": 0.871, "grad_norm": 0.21895843744277954, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51558426.0, "reward": 0.5611721873283386, "reward_std": 0.46151071786880493, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4973437488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.234375, "step": 4355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 895.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 704.375, "completions/mean_terminated_length": 704.375, "completions/min_length": 542.0, "completions/min_terminated_length": 542.0, "epoch": 0.8712, "grad_norm": 0.07635732740163803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51573616.0, "reward": 0.6055448055267334, "reward_std": 0.23901356756687164, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14858749508857727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 528.6875, "completions/mean_terminated_length": 528.6875, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8714, "grad_norm": 0.1842290461063385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51586187.0, "reward": 0.8771879076957703, "reward_std": 0.17546051740646362, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7543749809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3687499761581421, "step": 4357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 402.3125, "completions/mean_terminated_length": 402.3125, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.8716, "grad_norm": 0.2680032253265381, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51596768.0, "reward": 1.1349854469299316, "reward_std": 0.4796214997768402, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6449687480926514, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418750405311584, "step": 4358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1010.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 534.5625, "completions/mean_terminated_length": 534.5625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.8718, "grad_norm": 0.40083107352256775, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51609305.0, "reward": 0.79029780626297, "reward_std": 0.42299559712409973, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3305937647819519, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 4359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 767.625, "completions/mean_terminated_length": 731.0000610351562, "completions/min_length": 477.0, "completions/min_terminated_length": 477.0, "epoch": 0.872, "grad_norm": 0.22256730496883392, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51626051.0, "reward": 0.8388289213180542, "reward_std": 0.4934481382369995, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5526562333106995, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6156250238418579, "step": 4360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 888.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 569.9375, "completions/mean_terminated_length": 569.9375, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.8722, "grad_norm": 0.19503137469291687, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51638626.0, "reward": 0.7938759326934814, "reward_std": 0.42223408818244934, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33774998784065247, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "step": 4361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 257.125, "completions/mean_terminated_length": 257.125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.8724, "grad_norm": 0.3613903224468231, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51647076.0, "reward": 1.212188482284546, "reward_std": 0.39670127630233765, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.674375057220459, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7687499523162842, "step": 4362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 454.3125, "completions/mean_terminated_length": 454.3125, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.8726, "grad_norm": 0.25703859329223633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51658873.0, "reward": 1.1834132671356201, "reward_std": 0.43404507637023926, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7418249845504761, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8849999904632568, "step": 4363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 575.5, "completions/mean_terminated_length": 575.5, "completions/min_length": 479.0, "completions/min_terminated_length": 479.0, "epoch": 0.8728, "grad_norm": 0.004675393458455801, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51672305.0, "reward": 0.5396603941917419, "reward_std": 0.0168045312166214, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07931874692440033, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 423.6875, "completions/mean_terminated_length": 423.6875, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.873, "grad_norm": 0.2114473581314087, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51682572.0, "reward": 1.3981196880340576, "reward_std": 0.2834039628505707, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9212374687194824, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9300000071525574, "step": 4365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 676.25, "completions/mean_terminated_length": 676.25, "completions/min_length": 532.0, "completions/min_terminated_length": 532.0, "epoch": 0.8732, "grad_norm": 0.00020705517090391368, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51697568.0, "reward": 1.4990789890289307, "reward_std": 0.0005026470171287656, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981562495231628, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 426.625, "completions/mean_terminated_length": 386.8000183105469, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.8734, "grad_norm": 0.4232082962989807, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51708250.0, "reward": 1.1657696962356567, "reward_std": 0.5252346992492676, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7065375447273254, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 4367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 518.5625, "completions/mean_terminated_length": 446.357177734375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.8736, "grad_norm": 0.2751955986022949, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51720339.0, "reward": 0.5561726093292236, "reward_std": 0.272103488445282, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23734375834465027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6968749761581421, "step": 4368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 798.25, "completions/mean_terminated_length": 783.2000122070312, "completions/min_length": 564.0, "completions/min_terminated_length": 564.0, "epoch": 0.8738, "grad_norm": 0.1396641731262207, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51737943.0, "reward": 0.8071099519729614, "reward_std": 0.30129605531692505, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6767187118530273, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.34062498807907104, "step": 4369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 431.75, "completions/mean_terminated_length": 431.75, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.874, "grad_norm": 0.550918698310852, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51749187.0, "reward": 1.2019416093826294, "reward_std": 0.45482364296913147, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.716381311416626, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 335.6875, "completions/mean_terminated_length": 335.6875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.8742, "grad_norm": 0.2961592972278595, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51758462.0, "reward": 0.9213008880615234, "reward_std": 0.4671885669231415, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.467600017786026, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 4371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 380.4375, "completions/mean_terminated_length": 380.4375, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.8744, "grad_norm": 0.14777199923992157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51768581.0, "reward": 1.351850986480713, "reward_std": 0.32891175150871277, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8912000060081482, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.89124995470047, "step": 4372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 266.0, "completions/max_terminated_length": 266.0, "completions/mean_length": 211.625, "completions/mean_terminated_length": 211.625, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.8746, "grad_norm": 0.0007947339327074587, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51775775.0, "reward": 1.4994165897369385, "reward_std": 0.000607405963819474, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9988312721252441, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971875011920929, "step": 4373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 851.0, "completions/max_terminated_length": 851.0, "completions/mean_length": 474.3125, "completions/mean_terminated_length": 474.3125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.8748, "grad_norm": 0.3881261646747589, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51786980.0, "reward": 0.8111727237701416, "reward_std": 0.4539211690425873, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4348437488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.796875, "step": 4374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 547.6875, "completions/mean_terminated_length": 547.6875, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.875, "grad_norm": 0.26753807067871094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51801711.0, "reward": 0.8464727401733398, "reward_std": 0.45388343930244446, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3804437518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 635.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 450.875, "completions/mean_terminated_length": 450.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.8752, "grad_norm": 0.16874098777770996, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51812589.0, "reward": 0.6391415596008301, "reward_std": 0.24591480195522308, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21578124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 587.0625, "completions/mean_terminated_length": 524.6428833007812, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.8754, "grad_norm": 0.2091004103422165, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51825742.0, "reward": 0.7065475583076477, "reward_std": 0.43299296498298645, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4130937457084656, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6168749928474426, "step": 4377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 258.5, "completions/mean_terminated_length": 258.5, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.8756, "grad_norm": 0.5128586292266846, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51834406.0, "reward": 1.2363133430480957, "reward_std": 0.3968469202518463, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7851250171661377, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8450000286102295, "step": 4378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 453.625, "completions/mean_terminated_length": 453.625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.8758, "grad_norm": 0.017405018210411072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51845760.0, "reward": 0.5467853546142578, "reward_std": 0.01878262124955654, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09356874227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 478.8125, "completions/mean_terminated_length": 478.8125, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.876, "grad_norm": 0.14896926283836365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51857933.0, "reward": 1.46726655960083, "reward_std": 0.12593770027160645, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9970312118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 346.6875, "completions/mean_terminated_length": 346.6875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.8762, "grad_norm": 0.3384254574775696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51867384.0, "reward": 1.3335821628570557, "reward_std": 0.35792678594589233, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8546625375747681, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 4381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 954.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 498.0625, "completions/mean_terminated_length": 498.0625, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.8764, "grad_norm": 0.10156932473182678, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51880089.0, "reward": 0.5908572673797607, "reward_std": 0.1303544044494629, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18171249330043793, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8774999976158142, "step": 4382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 548.75, "completions/mean_terminated_length": 390.3333435058594, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.8766, "grad_norm": 0.37577521800994873, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51892981.0, "reward": 0.7398195862770081, "reward_std": 0.5970711708068848, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4171375036239624, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6599999666213989, "step": 4383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 512.75, "completions/mean_terminated_length": 512.75, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.8768, "grad_norm": 0.17913158237934113, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51905329.0, "reward": 0.7447509765625, "reward_std": 0.3836056590080261, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3019999861717224, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262500405311584, "step": 4384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 321.5, "completions/mean_terminated_length": 321.5, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.877, "grad_norm": 0.3127800524234772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51914425.0, "reward": 1.2645137310028076, "reward_std": 0.42076626420021057, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7790247201919556, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9661250114440918, "step": 4385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 716.4375, "completions/mean_terminated_length": 695.933349609375, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.8772, "grad_norm": 0.2529222071170807, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51929504.0, "reward": 0.6845697164535522, "reward_std": 0.4259902238845825, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24413751065731049, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 4386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 599.8125, "completions/mean_terminated_length": 501.923095703125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.8774, "grad_norm": 0.37466347217559814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51943053.0, "reward": 1.0069538354873657, "reward_std": 0.6242622137069702, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6389063000679016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 4387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 934.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 604.125, "completions/mean_terminated_length": 604.125, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.8776, "grad_norm": 0.18790893256664276, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51957647.0, "reward": 0.7139071226119995, "reward_std": 0.304575651884079, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30281251668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.824999988079071, "step": 4388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 620.0625, "completions/mean_terminated_length": 620.0625, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.8778, "grad_norm": 0.055523402988910675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51971920.0, "reward": 0.6991413831710815, "reward_std": 0.11247921735048294, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3982812762260437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7593750357627869, "step": 4389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 832.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 668.5, "completions/mean_terminated_length": 668.5, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.878, "grad_norm": 0.08305710554122925, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51987400.0, "reward": 0.5982665419578552, "reward_std": 0.1335620880126953, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1965312510728836, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8731249570846558, "step": 4390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 376.1875, "completions/mean_terminated_length": 376.1875, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.8782, "grad_norm": 0.4492063522338867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51997899.0, "reward": 1.0237510204315186, "reward_std": 0.49057790637016296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 489.0, "completions/mean_terminated_length": 489.0, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.8784, "grad_norm": 0.37840089201927185, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52010075.0, "reward": 1.377979040145874, "reward_std": 0.33134663105010986, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8809562921524048, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 4392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 581.375, "completions/mean_terminated_length": 581.375, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.8786, "grad_norm": 0.04372670501470566, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52023249.0, "reward": 0.5674228072166443, "reward_std": 0.05243338644504547, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13484373688697815, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 4393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 860.0, "completions/mean_length": 748.1875, "completions/mean_terminated_length": 656.25, "completions/min_length": 562.0, "completions/min_terminated_length": 562.0, "epoch": 0.8788, "grad_norm": 0.22422847151756287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52039156.0, "reward": 0.7305006384849548, "reward_std": 0.5998534560203552, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3985000252723694, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6725000143051147, "step": 4394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 361.1875, "completions/mean_terminated_length": 361.1875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.879, "grad_norm": 0.0006959181046113372, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52049015.0, "reward": 1.4994012117385864, "reward_std": 0.0005969053017906845, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987999796867371, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 4395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 234.5, "completions/mean_terminated_length": 234.5, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.8792, "grad_norm": 0.9788801670074463, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52056543.0, "reward": 1.2548633813858032, "reward_std": 0.43700504302978516, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.759725034236908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 4396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 376.125, "completions/mean_terminated_length": 376.125, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.8794, "grad_norm": 0.0003929673111997545, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52066465.0, "reward": 1.4990227222442627, "reward_std": 0.00048750487621873617, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980437755584717, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 981.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 701.625, "completions/mean_terminated_length": 701.625, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.8796, "grad_norm": 0.22800584137439728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52081931.0, "reward": 0.7085540294647217, "reward_std": 0.3346586525440216, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2921062409877777, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8768749833106995, "step": 4398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 378.1875, "completions/mean_terminated_length": 378.1875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.8798, "grad_norm": 0.3181888461112976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52092030.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 956.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 680.8125, "completions/mean_terminated_length": 680.8125, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.88, "grad_norm": 0.21748359501361847, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52107691.0, "reward": 0.6889073252677917, "reward_std": 0.335856556892395, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2528125047683716, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 4400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 419.625, "completions/mean_terminated_length": 419.625, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.8802, "grad_norm": 0.22333359718322754, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52118421.0, "reward": 0.6374103426933289, "reward_std": 0.25603851675987244, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21231874823570251, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 4401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 734.4375, "completions/mean_terminated_length": 693.0714721679688, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.8804, "grad_norm": 0.22902163863182068, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52134972.0, "reward": 1.2525668144226074, "reward_std": 0.543158233165741, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8176312446594238, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8518750667572021, "step": 4402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 404.8125, "completions/mean_terminated_length": 404.8125, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 0.8806, "grad_norm": 0.43958625197410583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52145081.0, "reward": 0.5871075987815857, "reward_std": 0.3059145510196686, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1742125004529953, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 4403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 689.875, "completions/mean_terminated_length": 667.6000366210938, "completions/min_length": 493.0, "completions/min_terminated_length": 493.0, "epoch": 0.8808, "grad_norm": 0.18919603526592255, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52160535.0, "reward": 0.6185696125030518, "reward_std": 0.3093288242816925, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23713751137256622, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8162499666213989, "step": 4404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 908.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 675.4375, "completions/mean_terminated_length": 675.4375, "completions/min_length": 478.0, "completions/min_terminated_length": 478.0, "epoch": 0.881, "grad_norm": 0.3098926842212677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52175454.0, "reward": 0.698279082775116, "reward_std": 0.27578720450401306, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3340562582015991, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7931250333786011, "step": 4405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 480.125, "completions/mean_terminated_length": 480.125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.8812, "grad_norm": 0.36888033151626587, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52187472.0, "reward": 0.8121883869171143, "reward_std": 0.4165906608104706, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37437501549720764, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 4406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 688.125, "completions/mean_terminated_length": 576.1666870117188, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.8814, "grad_norm": 0.3607226610183716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52202818.0, "reward": 0.6773444414138794, "reward_std": 0.5607796907424927, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.35468751192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6687499284744263, "step": 4407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 497.1875, "completions/mean_terminated_length": 497.1875, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.8816, "grad_norm": 0.6453251838684082, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52214837.0, "reward": 1.0005477666854858, "reward_std": 0.462074339389801, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5635937452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 432.75, "completions/mean_terminated_length": 432.75, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.8818, "grad_norm": 0.1294984370470047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52225729.0, "reward": 0.6287508606910706, "reward_std": 0.11485303193330765, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2574999928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 4409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 852.75, "completions/mean_terminated_length": 795.6666870117188, "completions/min_length": 662.0, "completions/min_terminated_length": 662.0, "epoch": 0.882, "grad_norm": 0.6805936694145203, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52243405.0, "reward": 0.5983597040176392, "reward_std": 0.43628424406051636, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5092187523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.2906249761581421, "step": 4410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 207.0, "completions/mean_terminated_length": 207.0, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 0.8822, "grad_norm": 0.759434163570404, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52250509.0, "reward": 1.1904137134552002, "reward_std": 0.4722740948200226, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6933245062828064, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9757499694824219, "step": 4411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 281.3125, "completions/mean_terminated_length": 281.3125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.8824, "grad_norm": 0.33861976861953735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52259138.0, "reward": 1.142707109451294, "reward_std": 0.4751651883125305, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6604125499725342, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 301.1875, "completions/mean_terminated_length": 301.1875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.8826, "grad_norm": 0.0002987584739457816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52267589.0, "reward": 1.4989478588104248, "reward_std": 0.00042326669790782034, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.997893750667572, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 434.9375, "completions/mean_terminated_length": 434.9375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.8828, "grad_norm": 0.262330025434494, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52278180.0, "reward": 1.1423447132110596, "reward_std": 0.4748761057853699, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6596875190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 508.75, "completions/mean_terminated_length": 474.4000244140625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.883, "grad_norm": 0.066715307533741, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52290592.0, "reward": 0.51734459400177, "reward_std": 0.1384401023387909, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09718749672174454, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 849.625, "completions/mean_terminated_length": 714.0, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.8832, "grad_norm": 0.27682632207870483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52307834.0, "reward": 0.5578129291534424, "reward_std": 0.5886431336402893, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3656249940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.42499998211860657, "step": 4416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 241.0625, "completions/mean_terminated_length": 241.0625, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.8834, "grad_norm": 0.20014852285385132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52315259.0, "reward": 0.6452352404594421, "reward_std": 0.13560038805007935, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2904687523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8156249523162842, "step": 4417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 332.125, "completions/mean_terminated_length": 332.125, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.8836, "grad_norm": 0.5687307119369507, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52324509.0, "reward": 0.7973321676254272, "reward_std": 0.42148110270500183, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.34466248750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "step": 4418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 328.6875, "completions/mean_terminated_length": 328.6875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.8838, "grad_norm": 0.34314095973968506, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52333192.0, "reward": 0.9208571910858154, "reward_std": 0.46522757411003113, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.46671247482299805, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9337500333786011, "step": 4419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 415.125, "completions/mean_terminated_length": 374.5333557128906, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.884, "grad_norm": 0.257723331451416, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52344010.0, "reward": 0.5909383893013, "reward_std": 0.2917596101760864, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18187499046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.862500011920929, "step": 4420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 350.25, "completions/mean_terminated_length": 350.25, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.8842, "grad_norm": 0.35313886404037476, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52353774.0, "reward": 0.804297685623169, "reward_std": 0.2232827991247177, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.608593761920929, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4593750238418579, "step": 4421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 682.375, "completions/mean_terminated_length": 659.6000366210938, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.8844, "grad_norm": 0.32501235604286194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52368532.0, "reward": 1.1202352046966553, "reward_std": 0.5544840693473816, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7404687404632568, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.828125, "step": 4422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 817.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 545.3125, "completions/mean_terminated_length": 545.3125, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.8846, "grad_norm": 0.2901197373867035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52381561.0, "reward": 0.9633476138114929, "reward_std": 0.3579925000667572, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6141937375068665, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6331249475479126, "step": 4423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 360.0, "completions/max_terminated_length": 360.0, "completions/mean_length": 278.875, "completions/mean_terminated_length": 278.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.8848, "grad_norm": 0.00035303906770423055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52389719.0, "reward": 1.4988821744918823, "reward_std": 0.0003586417296901345, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 888.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 510.5625, "completions/mean_terminated_length": 510.5625, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.885, "grad_norm": 0.42975887656211853, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52401824.0, "reward": 0.866172730922699, "reward_std": 0.4417284429073334, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41984373331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 818.1875, "completions/mean_terminated_length": 770.6923217773438, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.8852, "grad_norm": 0.4756353795528412, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52419603.0, "reward": 0.7781252861022949, "reward_std": 0.38919535279273987, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7437499761581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.1875, "step": 4426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 413.5, "completions/mean_terminated_length": 413.5, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.8854, "grad_norm": 0.30600738525390625, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52430091.0, "reward": 1.3773853778839111, "reward_std": 0.33309996128082275, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8797687292098999, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "step": 4427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 363.9375, "completions/mean_terminated_length": 363.9375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.8856, "grad_norm": 0.37092044949531555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52440250.0, "reward": 1.1737383604049683, "reward_std": 0.414760023355484, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.722474992275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7799999713897705, "step": 4428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 951.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 574.0, "completions/mean_terminated_length": 574.0, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.8858, "grad_norm": 0.20801810920238495, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52454778.0, "reward": 0.7450664639472961, "reward_std": 0.32153674960136414, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3651312589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8393750190734863, "step": 4429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 808.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 584.4375, "completions/mean_terminated_length": 584.4375, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.886, "grad_norm": 0.14625431597232819, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52470049.0, "reward": 0.7857007384300232, "reward_std": 0.24061821401119232, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5714000463485718, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4962500035762787, "step": 4430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 348.5625, "completions/mean_terminated_length": 348.5625, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.8862, "grad_norm": 0.3952538073062897, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52479194.0, "reward": 1.4391415119171143, "reward_std": 0.2374393194913864, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9407812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 382.0, "completions/max_terminated_length": 382.0, "completions/mean_length": 267.375, "completions/mean_terminated_length": 267.375, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.8864, "grad_norm": 0.9809848666191101, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52487584.0, "reward": 1.0643699169158936, "reward_std": 0.4886012077331543, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5662375092506409, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "step": 4432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 401.75, "completions/mean_terminated_length": 401.75, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.8866, "grad_norm": 0.3321405053138733, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52497404.0, "reward": 0.9050009250640869, "reward_std": 0.47499996423721313, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4350000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 584.0625, "completions/mean_terminated_length": 584.0625, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.8868, "grad_norm": 0.20016980171203613, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52510781.0, "reward": 1.142410397529602, "reward_std": 0.47557732462882996, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6598187685012817, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 340.1875, "completions/mean_terminated_length": 340.1875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.887, "grad_norm": 0.36625438928604126, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52519904.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 532.1875, "completions/mean_terminated_length": 532.1875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.8872, "grad_norm": 0.009657477959990501, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52534419.0, "reward": 0.9871876239776611, "reward_std": 0.013900639489293098, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9743750095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.14375001192092896, "step": 4436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 870.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 471.375, "completions/mean_terminated_length": 471.375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.8874, "grad_norm": 0.15465544164180756, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52546665.0, "reward": 0.5749887228012085, "reward_std": 0.14482298493385315, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1499750018119812, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8912500143051147, "step": 4437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 336.1875, "completions/mean_terminated_length": 336.1875, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.8876, "grad_norm": 0.39511242508888245, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52556092.0, "reward": 1.0097661018371582, "reward_std": 0.32258883118629456, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.70703125, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.559374988079071, "step": 4438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 267.1875, "completions/mean_terminated_length": 267.1875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.8878, "grad_norm": 0.0006607600371353328, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52564143.0, "reward": 1.499648094177246, "reward_std": 0.0005385674885474145, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9992937445640564, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "step": 4439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 463.875, "completions/mean_terminated_length": 463.875, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.888, "grad_norm": 0.6047573685646057, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52575453.0, "reward": 1.0392322540283203, "reward_std": 0.5526498556137085, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5784624814987183, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9649999737739563, "step": 4440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 355.0, "completions/mean_terminated_length": 355.0, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.8882, "grad_norm": 0.37653836607933044, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52585325.0, "reward": 1.2562229633331299, "reward_std": 0.43505677580833435, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7624437808990479, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 4441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 606.5625, "completions/mean_terminated_length": 606.5625, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.8884, "grad_norm": 0.10179904103279114, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52599494.0, "reward": 0.6078101992607117, "reward_std": 0.14767099916934967, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21561875939369202, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8493750095367432, "step": 4442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 281.875, "completions/mean_terminated_length": 281.875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.8886, "grad_norm": 0.39492931962013245, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52607972.0, "reward": 1.3815476894378662, "reward_std": 0.2529875636100769, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9505937099456787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8168749809265137, "step": 4443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 412.0, "completions/max_terminated_length": 412.0, "completions/mean_length": 329.8125, "completions/mean_terminated_length": 329.8125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.8888, "grad_norm": 0.237855464220047, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52617409.0, "reward": 1.4398071765899658, "reward_std": 0.2376156747341156, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9421125650405884, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 327.6875, "completions/mean_terminated_length": 327.6875, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.889, "grad_norm": 0.00045738479821011424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52626844.0, "reward": 1.4991540908813477, "reward_std": 0.0005386699340306222, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983062744140625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 819.9375, "completions/mean_terminated_length": 697.5, "completions/min_length": 586.0, "completions/min_terminated_length": 586.0, "epoch": 0.8892, "grad_norm": 0.2880074977874756, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52644587.0, "reward": 0.34631627798080444, "reward_std": 0.2794673442840576, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06763124465942383, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5893750190734863, "step": 4446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 459.25, "completions/mean_terminated_length": 459.25, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.8894, "grad_norm": 0.3203001320362091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52656639.0, "reward": 0.7966945171356201, "reward_std": 0.3803219795227051, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4058874845504761, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.82874995470047, "step": 4447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 484.4375, "completions/mean_terminated_length": 448.4666748046875, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.8896, "grad_norm": 0.3042559027671814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52668598.0, "reward": 1.1059198379516602, "reward_std": 0.5400115847587585, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6493375301361084, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 4448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 612.25, "completions/mean_terminated_length": 517.2307739257812, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.8898, "grad_norm": 0.8219427466392517, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52682266.0, "reward": 1.0003819465637207, "reward_std": 0.6002049446105957, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6882625222206116, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6449999809265137, "step": 4449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 689.0625, "completions/mean_terminated_length": 666.7333984375, "completions/min_length": 431.0, "completions/min_terminated_length": 431.0, "epoch": 0.89, "grad_norm": 0.26593631505966187, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52697819.0, "reward": 0.8926570415496826, "reward_std": 0.47072306275367737, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5353125333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 706.6875, "completions/mean_terminated_length": 600.9166870117188, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.8902, "grad_norm": 0.3534776270389557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52713046.0, "reward": 0.7486729025840759, "reward_std": 0.6040530800819397, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4348437488079071, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.609375, "step": 4451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 556.8125, "completions/mean_terminated_length": 525.6666870117188, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.8904, "grad_norm": 0.4269713759422302, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52725907.0, "reward": 1.283435344696045, "reward_std": 0.440036803483963, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8793687224388123, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7743749618530273, "step": 4452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 343.8125, "completions/mean_terminated_length": 343.8125, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.8906, "grad_norm": 0.29895317554473877, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52736624.0, "reward": 1.140225887298584, "reward_std": 0.47844091057777405, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6554499864578247, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 742.5, "completions/mean_terminated_length": 614.5454711914062, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.8908, "grad_norm": 0.17187750339508057, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52752200.0, "reward": 0.4889849126338959, "reward_std": 0.3876461982727051, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2904687523841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.43437498807907104, "step": 4454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 640.1875, "completions/mean_terminated_length": 585.357177734375, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.891, "grad_norm": 0.32747671008110046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52766891.0, "reward": 1.0695321559906006, "reward_std": 0.5639277100563049, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7015625238418579, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.731249988079071, "step": 4455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 668.25, "completions/mean_terminated_length": 668.25, "completions/min_length": 500.0, "completions/min_terminated_length": 500.0, "epoch": 0.8912, "grad_norm": 0.07873231917619705, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52782927.0, "reward": 0.5981133580207825, "reward_std": 0.1340566873550415, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19622500240802765, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8675000071525574, "step": 4456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 411.0, "completions/max_terminated_length": 411.0, "completions/mean_length": 333.125, "completions/mean_terminated_length": 333.125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.8914, "grad_norm": 0.4346866011619568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52792209.0, "reward": 1.1454572677612305, "reward_std": 0.47132450342178345, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6659125089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 340.3125, "completions/mean_terminated_length": 340.3125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.8916, "grad_norm": 0.7538538575172424, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52802422.0, "reward": 0.7685918807983398, "reward_std": 0.4358980357646942, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2871812582015991, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 4458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 324.0, "completions/mean_terminated_length": 324.0, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.8918, "grad_norm": 0.363890141248703, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52812038.0, "reward": 1.142650842666626, "reward_std": 0.4751201570034027, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6603000164031982, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "step": 4459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 572.0, "completions/mean_terminated_length": 541.86669921875, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.892, "grad_norm": 0.26277515292167664, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52825366.0, "reward": 1.3442041873931885, "reward_std": 0.4337778687477112, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.875906229019165, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9043750166893005, "step": 4460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 639.5625, "completions/mean_terminated_length": 613.933349609375, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.8922, "grad_norm": 0.23162131011486053, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52839615.0, "reward": 0.6360945701599121, "reward_std": 0.3638868033885956, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2096875011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 386.5625, "completions/mean_terminated_length": 386.5625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.8924, "grad_norm": 0.44190460443496704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52849656.0, "reward": 1.1989073753356934, "reward_std": 0.45963767170906067, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7103124856948853, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 4462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 996.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 581.6875, "completions/mean_terminated_length": 581.6875, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.8926, "grad_norm": 0.06077095866203308, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52863507.0, "reward": 0.6178789734840393, "reward_std": 0.1046692281961441, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23575624823570251, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8643749952316284, "step": 4463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 626.125, "completions/mean_terminated_length": 626.125, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.8928, "grad_norm": 0.2305528223514557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52877285.0, "reward": 0.8585945963859558, "reward_std": 0.38814103603363037, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4671875238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 4464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 628.375, "completions/mean_terminated_length": 602.0000610351562, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.893, "grad_norm": 0.5853680372238159, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52891691.0, "reward": 0.79007887840271, "reward_std": 0.4081370234489441, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.45515623688697815, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7531249523162842, "step": 4465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 529.75, "completions/mean_terminated_length": 496.8000183105469, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.8932, "grad_norm": 0.3199022710323334, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52903831.0, "reward": 1.0486602783203125, "reward_std": 0.5428298115730286, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5973187685012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8893749713897705, "step": 4466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 319.875, "completions/mean_terminated_length": 319.875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.8934, "grad_norm": 0.0003016532864421606, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52913893.0, "reward": 1.4989665746688843, "reward_std": 0.00046470126835629344, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979312419891357, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "step": 4467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 502.9375, "completions/mean_terminated_length": 428.5000305175781, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.8936, "grad_norm": 0.5498798489570618, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52925908.0, "reward": 1.0723446607589722, "reward_std": 0.5889527797698975, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6446875333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 4468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 589.6875, "completions/mean_terminated_length": 560.7333374023438, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.8938, "grad_norm": 0.18454957008361816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52939071.0, "reward": 0.6121101975440979, "reward_std": 0.2974131107330322, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22421875596046448, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.828125, "step": 4469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 909.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 688.625, "completions/mean_terminated_length": 688.625, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.894, "grad_norm": 0.19745789468288422, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52954089.0, "reward": 1.2048447132110596, "reward_std": 0.4505309760570526, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7221875190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 679.6875, "completions/mean_terminated_length": 656.7333374023438, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.8942, "grad_norm": 0.16290293633937836, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52968628.0, "reward": 0.8729690313339233, "reward_std": 0.2725021541118622, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8084374666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.23749999701976776, "step": 4471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 308.8125, "completions/mean_terminated_length": 308.8125, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.8944, "grad_norm": 0.6596897840499878, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52977553.0, "reward": 0.8644227385520935, "reward_std": 0.4537990987300873, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4163437485694885, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9181250333786011, "step": 4472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 605.5625, "completions/mean_terminated_length": 577.6666870117188, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.8946, "grad_norm": 0.19306300580501556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52991226.0, "reward": 0.9946072101593018, "reward_std": 0.5399571061134338, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5517125129699707, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8837499618530273, "step": 4473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 415.5, "completions/mean_terminated_length": 415.5, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.8948, "grad_norm": 0.421988308429718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53002306.0, "reward": 1.320625901222229, "reward_std": 0.38295724987983704, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8287500143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 698.875, "completions/mean_terminated_length": 652.4285888671875, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.895, "grad_norm": 0.1374415010213852, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53017344.0, "reward": 0.624219536781311, "reward_std": 0.3495098352432251, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3109374940395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 403.0, "completions/mean_terminated_length": 403.0, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.8952, "grad_norm": 0.012025329284369946, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53028336.0, "reward": 0.5492165684700012, "reward_std": 0.015579965896904469, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09843124449253082, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 4476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 679.125, "completions/mean_terminated_length": 629.857177734375, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.8954, "grad_norm": 0.2653409540653229, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53043346.0, "reward": 0.6884381771087646, "reward_std": 0.34903380274772644, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.43937498331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.637499988079071, "step": 4477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 258.0, "completions/max_terminated_length": 258.0, "completions/mean_length": 202.625, "completions/mean_terminated_length": 202.625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.8956, "grad_norm": 0.5580751895904541, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53050412.0, "reward": 0.9552167654037476, "reward_std": 0.4957176744937897, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47293126583099365, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "step": 4478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 535.5, "completions/mean_terminated_length": 535.5, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.8958, "grad_norm": 0.20571424067020416, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53062548.0, "reward": 1.1417853832244873, "reward_std": 0.43967926502227783, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7210687398910522, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8331249952316284, "step": 4479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 594.4375, "completions/mean_terminated_length": 565.800048828125, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.896, "grad_norm": 0.2053300440311432, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53077003.0, "reward": 1.2850632667541504, "reward_std": 0.47444769740104675, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8201249837875366, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9087499976158142, "step": 4480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 356.6875, "completions/mean_terminated_length": 356.6875, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.8962, "grad_norm": 0.14148810505867004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53086598.0, "reward": 0.5954228639602661, "reward_std": 0.24147853255271912, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12834374606609344, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 534.6875, "completions/mean_terminated_length": 534.6875, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.8964, "grad_norm": 0.10834360122680664, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53098993.0, "reward": 1.438204050064087, "reward_std": 0.23723359405994415, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9389062523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 4482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 565.25, "completions/mean_terminated_length": 565.25, "completions/min_length": 409.0, "completions/min_terminated_length": 409.0, "epoch": 0.8966, "grad_norm": 0.2204275131225586, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53111813.0, "reward": 0.9120321273803711, "reward_std": 0.41852253675460815, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5115624666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.78125, "step": 4483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 735.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 561.375, "completions/mean_terminated_length": 561.375, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.8968, "grad_norm": 0.19513384997844696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53124699.0, "reward": 1.0324227809906006, "reward_std": 0.4401397705078125, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6273437738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8468749523162842, "step": 4484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 354.625, "completions/mean_terminated_length": 354.625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.897, "grad_norm": 0.2233751267194748, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53134117.0, "reward": 0.7217824459075928, "reward_std": 0.3879466652870178, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25606250762939453, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 449.125, "completions/mean_terminated_length": 449.125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.8972, "grad_norm": 0.0015826268354430795, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53146119.0, "reward": 0.9950002431869507, "reward_std": 0.0044723027385771275, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9900000095367432, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.08750000596046448, "step": 4486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 807.0, "completions/max_terminated_length": 807.0, "completions/mean_length": 459.6875, "completions/mean_terminated_length": 459.6875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.8974, "grad_norm": 0.05741271376609802, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53157362.0, "reward": 0.6015509366989136, "reward_std": 0.12474966794252396, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20309999585151672, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8612499833106995, "step": 4487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 475.3125, "completions/mean_terminated_length": 475.3125, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.8976, "grad_norm": 0.05094846710562706, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53169415.0, "reward": 0.6519540548324585, "reward_std": 0.24114388227462769, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.24140624701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968749642372131, "step": 4488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 475.0, "completions/max_terminated_length": 475.0, "completions/mean_length": 317.0625, "completions/mean_terminated_length": 317.0625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.8978, "grad_norm": 0.26055294275283813, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53178456.0, "reward": 1.1435822248458862, "reward_std": 0.4737643599510193, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6621625423431396, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 590.9375, "completions/mean_terminated_length": 562.0667114257812, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.898, "grad_norm": 0.22470706701278687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53191975.0, "reward": 0.8653290867805481, "reward_std": 0.5244666934013367, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.418156236410141, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9031250476837158, "step": 4490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 391.6875, "completions/mean_terminated_length": 391.6875, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.8982, "grad_norm": 0.17663511633872986, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53201778.0, "reward": 0.66382896900177, "reward_std": 0.25775718688964844, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.26515623927116394, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.846875011920929, "step": 4491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 323.25, "completions/mean_terminated_length": 323.25, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.8984, "grad_norm": 0.30501508712768555, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53210422.0, "reward": 1.1953572034835815, "reward_std": 0.46519044041633606, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7032124996185303, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 440.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 377.125, "completions/mean_terminated_length": 377.125, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.8986, "grad_norm": 0.2532907724380493, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53220408.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 397.125, "completions/mean_terminated_length": 397.125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.8988, "grad_norm": 0.13589459657669067, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53230794.0, "reward": 1.4372228384017944, "reward_std": 0.24727298319339752, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.936943769454956, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 4494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 287.8125, "completions/mean_terminated_length": 287.8125, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.899, "grad_norm": 0.00034328739275224507, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53239015.0, "reward": 1.499110460281372, "reward_std": 0.000552769866771996, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9982187747955322, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 851.0, "completions/mean_terminated_length": 811.0769653320312, "completions/min_length": 636.0, "completions/min_terminated_length": 636.0, "epoch": 0.8992, "grad_norm": 0.21133017539978027, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53256951.0, "reward": 0.8158474564552307, "reward_std": 0.5263673067092896, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5691937208175659, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5581250190734863, "step": 4496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 284.4375, "completions/mean_terminated_length": 284.4375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.8994, "grad_norm": 0.42884016036987305, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53265102.0, "reward": 0.7725573182106018, "reward_std": 0.4337700307369232, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2951124906539917, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "step": 4497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 382.0, "completions/max_terminated_length": 382.0, "completions/mean_length": 309.625, "completions/mean_terminated_length": 309.625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.8996, "grad_norm": 0.00024509162176400423, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53274072.0, "reward": 1.4988821744918823, "reward_std": 0.0003586417296901345, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 367.0, "completions/mean_terminated_length": 367.0, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.8998, "grad_norm": 0.13997454941272736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53283336.0, "reward": 0.6475417613983154, "reward_std": 0.3331114947795868, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1700812578201294, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "step": 4499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 440.0625, "completions/mean_terminated_length": 440.0625, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.9, "grad_norm": 0.2649633288383484, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53293817.0, "reward": 0.7491384744644165, "reward_std": 0.3892345428466797, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3107749819755554, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 4500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 315.75, "completions/mean_terminated_length": 315.75, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.9002, "grad_norm": 0.2725007236003876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53302245.0, "reward": 0.686719536781311, "reward_std": 0.17402441799640656, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3734374940395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.731249988079071, "step": 4501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 311.1875, "completions/mean_terminated_length": 311.1875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.9004, "grad_norm": 0.4385123550891876, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53310872.0, "reward": 0.7872477769851685, "reward_std": 0.42498570680618286, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32449376583099365, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 4502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 964.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 582.0, "completions/mean_terminated_length": 582.0, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.9006, "grad_norm": 0.09812003374099731, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53324760.0, "reward": 0.71070396900177, "reward_std": 0.23714660108089447, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35890626907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 4503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 548.3125, "completions/mean_terminated_length": 548.3125, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.9008, "grad_norm": 0.02484120987355709, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53338109.0, "reward": 0.5778133869171143, "reward_std": 0.08138053864240646, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15562500059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 4504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 295.125, "completions/mean_terminated_length": 295.125, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.901, "grad_norm": 0.7002626061439514, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53346831.0, "reward": 1.019726037979126, "reward_std": 0.4952867031097412, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5394500494003296, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 900.5625, "completions/mean_terminated_length": 804.5555419921875, "completions/min_length": 637.0, "completions/min_terminated_length": 637.0, "epoch": 0.9012, "grad_norm": 0.31232133507728577, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53365208.0, "reward": 0.630534827709198, "reward_std": 0.6439080238342285, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.44856876134872437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.3643749952316284, "step": 4506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 414.875, "completions/mean_terminated_length": 414.875, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.9014, "grad_norm": 0.12280549854040146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53375766.0, "reward": 0.8861722946166992, "reward_std": 0.19573181867599487, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7723437547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.30937498807907104, "step": 4507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 187.0, "completions/max_terminated_length": 187.0, "completions/mean_length": 153.6875, "completions/mean_terminated_length": 153.6875, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 0.9016, "grad_norm": 0.6723324656486511, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53381953.0, "reward": 0.7046480178833008, "reward_std": 0.39497730135917664, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2217937409877777, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 4508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 687.9375, "completions/mean_terminated_length": 665.5333862304688, "completions/min_length": 555.0, "completions/min_terminated_length": 555.0, "epoch": 0.9018, "grad_norm": 0.21891126036643982, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53396784.0, "reward": 1.062969446182251, "reward_std": 0.49136048555374146, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6884374618530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.793749988079071, "step": 4509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 856.1875, "completions/mean_terminated_length": 800.25, "completions/min_length": 566.0, "completions/min_terminated_length": 566.0, "epoch": 0.902, "grad_norm": 0.3009616732597351, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53414915.0, "reward": 0.5094537734985352, "reward_std": 0.3924572467803955, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20640623569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6468749642372131, "step": 4510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 521.875, "completions/mean_terminated_length": 521.875, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.9022, "grad_norm": 0.2124021351337433, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53427601.0, "reward": 0.8203821182250977, "reward_std": 0.421438604593277, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3907625079154968, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8887499570846558, "step": 4511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 788.9375, "completions/mean_terminated_length": 710.5833740234375, "completions/min_length": 543.0, "completions/min_terminated_length": 543.0, "epoch": 0.9024, "grad_norm": 0.20531751215457916, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53444112.0, "reward": 0.5419288873672485, "reward_std": 0.4507145583629608, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2088562548160553, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7006250023841858, "step": 4512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 546.5625, "completions/mean_terminated_length": 546.5625, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.9026, "grad_norm": 0.17253145575523376, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53456441.0, "reward": 1.0732821226119995, "reward_std": 0.44732803106307983, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6465624570846558, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8312499523162842, "step": 4513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 404.75, "completions/mean_terminated_length": 404.75, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.9028, "grad_norm": 0.29868829250335693, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53466757.0, "reward": 1.1948760747909546, "reward_std": 0.4662712812423706, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7022500038146973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "step": 4514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 337.8125, "completions/mean_terminated_length": 337.8125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.903, "grad_norm": 0.45300954580307007, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53476082.0, "reward": 1.3336477279663086, "reward_std": 0.35795921087265015, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8547937870025635, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9381250143051147, "step": 4515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 561.1875, "completions/mean_terminated_length": 561.1875, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.9032, "grad_norm": 0.06218115985393524, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53489845.0, "reward": 0.6250790357589722, "reward_std": 0.24260540306568146, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1876562535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 4516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 295.625, "completions/mean_terminated_length": 295.625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.9034, "grad_norm": 0.22778715193271637, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53498495.0, "reward": 1.3580100536346436, "reward_std": 0.2639644742012024, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.7785187363624573, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.768125057220459, "step": 4517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 430.3125, "completions/mean_terminated_length": 390.7333679199219, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.9036, "grad_norm": 0.19940650463104248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53509780.0, "reward": 1.2524758577346802, "reward_std": 0.4744469225406647, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8174500465393066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8537499904632568, "step": 4518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 513.4375, "completions/mean_terminated_length": 513.4375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.9038, "grad_norm": 0.19894090294837952, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53522203.0, "reward": 0.7676571011543274, "reward_std": 0.38144102692604065, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3478124737739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 4519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 772.5625, "completions/mean_terminated_length": 736.6428833007812, "completions/min_length": 533.0, "completions/min_terminated_length": 533.0, "epoch": 0.904, "grad_norm": 0.18389885127544403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53538340.0, "reward": 0.7089067697525024, "reward_std": 0.305416077375412, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5428124666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5, "step": 4520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 822.0, "completions/max_terminated_length": 822.0, "completions/mean_length": 547.0, "completions/mean_terminated_length": 547.0, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.9042, "grad_norm": 0.220628023147583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53550868.0, "reward": 1.0293757915496826, "reward_std": 0.47475865483283997, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5587500333786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 4521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 673.0, "completions/mean_terminated_length": 649.6000366210938, "completions/min_length": 520.0, "completions/min_terminated_length": 520.0, "epoch": 0.9044, "grad_norm": 0.321340411901474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53565668.0, "reward": 1.1290103197097778, "reward_std": 0.5201279520988464, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6955187320709229, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8756250143051147, "step": 4522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 480.3125, "completions/mean_terminated_length": 480.3125, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.9046, "grad_norm": 0.21640056371688843, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53577257.0, "reward": 0.6808352470397949, "reward_std": 0.32671210169792175, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23666873574256897, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9318749904632568, "step": 4523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 408.5625, "completions/mean_terminated_length": 408.5625, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.9048, "grad_norm": 0.21893343329429626, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53588466.0, "reward": 1.1892726421356201, "reward_std": 0.42798280715942383, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7535437941551208, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8818750381469727, "step": 4524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 487.75, "completions/mean_terminated_length": 487.75, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.905, "grad_norm": 0.3205988109111786, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53601870.0, "reward": 1.0860165357589722, "reward_std": 0.48347142338752747, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6095312237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 746.75, "completions/mean_terminated_length": 654.3333740234375, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.9052, "grad_norm": 0.2014622688293457, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53618650.0, "reward": 0.47288811206817627, "reward_std": 0.30978766083717346, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19577500224113464, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6212500333786011, "step": 4526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 323.375, "completions/mean_terminated_length": 323.375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.9054, "grad_norm": 0.0006846374017186463, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53627904.0, "reward": 1.499341607093811, "reward_std": 0.0006110992981120944, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9986813068389893, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 4527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 526.75, "completions/mean_terminated_length": 526.75, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.9056, "grad_norm": 0.11906138062477112, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53641868.0, "reward": 0.8042161464691162, "reward_std": 0.21684502065181732, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6084312200546265, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4868749678134918, "step": 4528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 284.0, "completions/mean_terminated_length": 284.0, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.9058, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53650204.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 311.0, "completions/max_terminated_length": 311.0, "completions/mean_length": 250.5625, "completions/mean_terminated_length": 250.5625, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.906, "grad_norm": 0.3458760678768158, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53657605.0, "reward": 0.7608106136322021, "reward_std": 0.4410724341869354, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.27161872386932373, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831249713897705, "step": 4530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 503.5625, "completions/mean_terminated_length": 503.5625, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.9062, "grad_norm": 0.32273420691490173, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53669566.0, "reward": 1.3787915706634521, "reward_std": 0.32970789074897766, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8825812339782715, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 4531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 455.0, "completions/mean_terminated_length": 455.0, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.9064, "grad_norm": 0.0916733592748642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53681422.0, "reward": 0.7359381318092346, "reward_std": 0.18978433310985565, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47187501192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.637499988079071, "step": 4532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 491.0625, "completions/mean_terminated_length": 455.5333557128906, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.9066, "grad_norm": 0.2697490155696869, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53692767.0, "reward": 0.7106133699417114, "reward_std": 0.42158353328704834, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29622501134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8675000071525574, "step": 4533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 699.0, "completions/max_terminated_length": 699.0, "completions/mean_length": 446.5625, "completions/mean_terminated_length": 446.5625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.9068, "grad_norm": 0.16760112345218658, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53703512.0, "reward": 1.2843726873397827, "reward_std": 0.38948825001716614, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8187437057495117, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9181250333786011, "step": 4534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 509.5625, "completions/mean_terminated_length": 475.2666931152344, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.907, "grad_norm": 0.4179861545562744, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53715361.0, "reward": 1.167578935623169, "reward_std": 0.5229057669639587, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7101563215255737, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 4535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 507.9375, "completions/mean_terminated_length": 507.9375, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.9072, "grad_norm": 0.10495465993881226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53727136.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 678.0, "completions/max_terminated_length": 678.0, "completions/mean_length": 417.5, "completions/mean_terminated_length": 417.5, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.9074, "grad_norm": 0.048661816865205765, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53738152.0, "reward": 0.5960946679115295, "reward_std": 0.08089964836835861, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19218748807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "step": 4537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 596.1875, "completions/mean_terminated_length": 567.6666870117188, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.9076, "grad_norm": 0.2782224118709564, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53751707.0, "reward": 1.190782070159912, "reward_std": 0.4704347252845764, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8190625309944153, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 4538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 389.8125, "completions/mean_terminated_length": 389.8125, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.9078, "grad_norm": 0.3643695116043091, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53761608.0, "reward": 1.0469539165496826, "reward_std": 0.4294934570789337, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6564062833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8343749642372131, "step": 4539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 696.0, "completions/mean_terminated_length": 649.1428833007812, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.908, "grad_norm": 0.28360432386398315, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53777416.0, "reward": 0.5938163995742798, "reward_std": 0.3536348044872284, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2501312494277954, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7206249833106995, "step": 4540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 917.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 522.6875, "completions/mean_terminated_length": 522.6875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.9082, "grad_norm": 0.1318187117576599, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53789459.0, "reward": 1.46726655960083, "reward_std": 0.12593770027160645, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9970312118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 279.0, "completions/max_terminated_length": 279.0, "completions/mean_length": 210.125, "completions/mean_terminated_length": 210.125, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.9084, "grad_norm": 0.0003973868442699313, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53796693.0, "reward": 1.4988915920257568, "reward_std": 0.00038525793934240937, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977812767028809, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 4542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 450.75, "completions/mean_terminated_length": 450.75, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.9086, "grad_norm": 0.21336455643177032, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53807681.0, "reward": 0.8635161519050598, "reward_std": 0.1928780972957611, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7270312309265137, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3843750059604645, "step": 4543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 587.3125, "completions/mean_terminated_length": 558.2000122070312, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.9088, "grad_norm": 0.1618366539478302, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53820774.0, "reward": 0.6332038640975952, "reward_std": 0.22349074482917786, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32890623807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 4544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 704.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 484.0625, "completions/mean_terminated_length": 484.0625, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.909, "grad_norm": 0.14273099601268768, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53833367.0, "reward": 0.7700758576393127, "reward_std": 0.37508389353752136, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3526500165462494, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "step": 4545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 560.0, "completions/max_terminated_length": 560.0, "completions/mean_length": 397.3125, "completions/mean_terminated_length": 397.3125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.9092, "grad_norm": 0.06397099047899246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53844412.0, "reward": 0.6044477224349976, "reward_std": 0.12249982357025146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20889374613761902, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8706250190734863, "step": 4546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 413.8125, "completions/mean_terminated_length": 413.8125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.9094, "grad_norm": 0.1739620715379715, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53855593.0, "reward": 1.0770316123962402, "reward_std": 0.20438335835933685, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.9665625095367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.2750000059604645, "step": 4547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 785.0, "completions/max_terminated_length": 785.0, "completions/mean_length": 510.0, "completions/mean_terminated_length": 510.0, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.9096, "grad_norm": 0.142586812376976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53867753.0, "reward": 0.9246882200241089, "reward_std": 0.3126085102558136, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6618750095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.668749988079071, "step": 4548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 508.1875, "completions/mean_terminated_length": 508.1875, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.9098, "grad_norm": 0.003697258187457919, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53879564.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 281.0, "completions/mean_terminated_length": 281.0, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.91, "grad_norm": 0.0003218957281205803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53889004.0, "reward": 1.4989571571350098, "reward_std": 0.00044457349576987326, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979125261306763, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 384.6875, "completions/mean_terminated_length": 384.6875, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.9102, "grad_norm": 0.2740963399410248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53899527.0, "reward": 1.2610915899276733, "reward_std": 0.42475906014442444, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7721812725067139, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "step": 4551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 397.5, "completions/mean_terminated_length": 397.5, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.9104, "grad_norm": 0.5007504820823669, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53909695.0, "reward": 0.8376790285110474, "reward_std": 0.4602908194065094, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3628562390804291, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 630.0625, "completions/mean_terminated_length": 630.0625, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.9106, "grad_norm": 0.10357999056577682, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53924032.0, "reward": 0.7940632700920105, "reward_std": 0.3144967257976532, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4631250202655792, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 4553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 284.0, "completions/max_terminated_length": 284.0, "completions/mean_length": 230.125, "completions/mean_terminated_length": 230.125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.9108, "grad_norm": 0.00029838093905709684, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53931666.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 194.375, "completions/mean_terminated_length": 194.375, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.911, "grad_norm": 0.37510186433792114, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53938440.0, "reward": 0.8440631031990051, "reward_std": 0.20058374106884003, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6881250143051147, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.41874998807907104, "step": 4555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 367.9375, "completions/mean_terminated_length": 367.9375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.9112, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53948183.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 787.3125, "completions/mean_terminated_length": 753.5000610351562, "completions/min_length": 504.0, "completions/min_terminated_length": 504.0, "epoch": 0.9114, "grad_norm": 0.12748903036117554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53964700.0, "reward": 0.5628913640975952, "reward_std": 0.2539459764957428, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25078123807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7218749523162842, "step": 4557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 435.4375, "completions/mean_terminated_length": 435.4375, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.9116, "grad_norm": 0.06303118169307709, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53976627.0, "reward": 1.4397227764129639, "reward_std": 0.23759308457374573, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419437646865845, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 4558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 522.5, "completions/mean_terminated_length": 522.5, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.9118, "grad_norm": 0.1908002346754074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53988955.0, "reward": 0.8403133153915405, "reward_std": 0.4059324264526367, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4306250214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 4559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 441.0666809082031, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.912, "grad_norm": 0.3221594989299774, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54000483.0, "reward": 1.2270853519439697, "reward_std": 0.5015257000923157, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7666687965393066, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8943749666213989, "step": 4560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 694.8125, "completions/mean_terminated_length": 647.7857666015625, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.9122, "grad_norm": 0.1304994374513626, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54015920.0, "reward": 0.7228507995605469, "reward_std": 0.46912339329719543, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3832000195980072, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7237499952316284, "step": 4561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 485.8125, "completions/mean_terminated_length": 485.8125, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.9124, "grad_norm": 0.18143294751644135, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54028029.0, "reward": 0.6553789377212524, "reward_std": 0.16537949442863464, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.31075623631477356, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7831250429153442, "step": 4562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 414.625, "completions/mean_terminated_length": 414.625, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.9126, "grad_norm": 0.4256903827190399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54038279.0, "reward": 0.9384258389472961, "reward_std": 0.4512406289577484, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5018500089645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8737499713897705, "step": 4563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 668.6875, "completions/mean_terminated_length": 645.0000610351562, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.9128, "grad_norm": 0.12625406682491302, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54053362.0, "reward": 0.6595320105552673, "reward_std": 0.2839968204498291, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3190625011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7875000238418579, "step": 4564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 321.3125, "completions/mean_terminated_length": 321.3125, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.913, "grad_norm": 0.12825152277946472, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54062343.0, "reward": 1.4400789737701416, "reward_std": 0.12348068505525589, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.8801562190055847, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8031250238418579, "step": 4565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 483.8125, "completions/mean_terminated_length": 483.8125, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.9132, "grad_norm": 0.197904571890831, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54074116.0, "reward": 0.7920321822166443, "reward_std": 0.4216875433921814, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33406251668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 671.625, "completions/mean_terminated_length": 648.1333618164062, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.9134, "grad_norm": 0.2590824067592621, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54089006.0, "reward": 0.7416411638259888, "reward_std": 0.29271048307418823, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5457812547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.46562501788139343, "step": 4567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 501.75, "completions/mean_terminated_length": 501.75, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.9136, "grad_norm": 0.3600507974624634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54101226.0, "reward": 1.212301254272461, "reward_std": 0.4477561116218567, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7371000051498413, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9450000524520874, "step": 4568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 707.3125, "completions/mean_terminated_length": 662.0714721679688, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.9138, "grad_norm": 0.2677086889743805, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54116543.0, "reward": 0.7646881341934204, "reward_std": 0.41861918568611145, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.40437501668930054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5, "step": 4569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 604.75, "completions/mean_terminated_length": 604.75, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.914, "grad_norm": 0.23768408596515656, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54130123.0, "reward": 0.9557820558547974, "reward_std": 0.439407080411911, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5365625023841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8125, "step": 4570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 406.1875, "completions/mean_terminated_length": 365.0000305175781, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.9142, "grad_norm": 0.45742493867874146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54140606.0, "reward": 0.8678103685379028, "reward_std": 0.5224545001983643, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4231187701225281, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8993749618530273, "step": 4571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 305.9375, "completions/mean_terminated_length": 305.9375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.9144, "grad_norm": 0.0006901623564772308, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54149165.0, "reward": 1.49879789352417, "reward_std": 0.0011288125533610582, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975937604904175, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 594.0625, "completions/mean_terminated_length": 594.0625, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.9146, "grad_norm": 0.047790784388780594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54163102.0, "reward": 0.6166415214538574, "reward_std": 0.11729899048805237, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.23328125476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "step": 4573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 595.9375, "completions/mean_terminated_length": 567.4000244140625, "completions/min_length": 456.0, "completions/min_terminated_length": 456.0, "epoch": 0.9148, "grad_norm": 0.14421594142913818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54177229.0, "reward": 0.9155475497245789, "reward_std": 0.4354000687599182, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6435937881469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.590624988079071, "step": 4574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1001.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 539.4375, "completions/mean_terminated_length": 539.4375, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.915, "grad_norm": 0.12248355895280838, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54189940.0, "reward": 0.5950790047645569, "reward_std": 0.10355482250452042, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1901562511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "step": 4575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 658.4375, "completions/mean_terminated_length": 574.0769653320312, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.9152, "grad_norm": 0.15116462111473083, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54205035.0, "reward": 0.5049100518226624, "reward_std": 0.28247976303100586, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19731874763965607, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6831250190734863, "step": 4576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 443.9375, "completions/mean_terminated_length": 443.9375, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.9154, "grad_norm": 0.10493998974561691, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54215962.0, "reward": 0.6282041668891907, "reward_std": 0.24724678695201874, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1939062476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 4577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 395.6875, "completions/mean_terminated_length": 395.6875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.9156, "grad_norm": 0.27632033824920654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54225973.0, "reward": 0.8153789639472961, "reward_std": 0.36604616045951843, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4432562589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8206250071525574, "step": 4578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 223.0, "completions/max_terminated_length": 223.0, "completions/mean_length": 169.75, "completions/mean_terminated_length": 169.75, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.9158, "grad_norm": 0.21427424252033234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54232769.0, "reward": 0.5940415859222412, "reward_std": 0.24253620207309723, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12558124959468842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 4579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 522.0, "completions/mean_terminated_length": 488.5333557128906, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.916, "grad_norm": 0.2316339761018753, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54245089.0, "reward": 1.3096632957458496, "reward_std": 0.3536255955696106, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8693249821662903, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.872499942779541, "step": 4580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 240.9375, "completions/mean_terminated_length": 240.9375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.9162, "grad_norm": 0.0005415870109573007, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54252976.0, "reward": 1.4988946914672852, "reward_std": 0.0003946831275243312, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977874755859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 4581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 465.8125, "completions/mean_terminated_length": 465.8125, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.9164, "grad_norm": 0.595607340335846, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54264541.0, "reward": 1.259441614151001, "reward_std": 0.4281385838985443, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7688812613487244, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 343.5, "completions/mean_terminated_length": 343.5, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.9166, "grad_norm": 0.3410973846912384, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54273973.0, "reward": 1.2867071628570557, "reward_std": 0.39019352197647095, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8234125375747681, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 4583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 390.4375, "completions/mean_terminated_length": 390.4375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.9168, "grad_norm": 0.18064755201339722, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54284492.0, "reward": 0.6081259250640869, "reward_std": 0.23749999701976776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1537500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 339.0, "completions/max_terminated_length": 339.0, "completions/mean_length": 281.1875, "completions/mean_terminated_length": 281.1875, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.917, "grad_norm": 0.42520859837532043, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54293391.0, "reward": 0.8960665464401245, "reward_std": 0.48269155621528625, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.417131245136261, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "step": 4585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 457.3125, "completions/mean_terminated_length": 457.3125, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.9172, "grad_norm": 0.33643126487731934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54304820.0, "reward": 1.019132375717163, "reward_std": 0.49597790837287903, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5382625460624695, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 555.125, "completions/mean_terminated_length": 555.125, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.9174, "grad_norm": 0.00020227003551553935, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54317830.0, "reward": 1.4990884065628052, "reward_std": 0.0005181764718145132, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9981750249862671, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "step": 4587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 477.5, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.9176, "grad_norm": 0.341991126537323, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54330286.0, "reward": 1.3335821628570557, "reward_std": 0.3550844192504883, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8546625375747681, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "step": 4588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 566.5625, "completions/mean_terminated_length": 566.5625, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.9178, "grad_norm": 0.29173851013183594, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54344791.0, "reward": 1.20750093460083, "reward_std": 0.4466449022293091, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7275000214576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 241.0, "completions/max_terminated_length": 241.0, "completions/mean_length": 201.375, "completions/mean_terminated_length": 201.375, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.918, "grad_norm": 0.5854265689849854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54351677.0, "reward": 0.7639541625976562, "reward_std": 0.43878498673439026, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.27790623903274536, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "step": 4590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 784.125, "completions/mean_terminated_length": 640.2000122070312, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.9182, "grad_norm": 0.3575124144554138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54368927.0, "reward": 0.711941123008728, "reward_std": 0.6502841711044312, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.48638126254081726, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.4831250309944153, "step": 4591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 487.0625, "completions/mean_terminated_length": 487.0625, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.9184, "grad_norm": 0.14655373990535736, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54380480.0, "reward": 1.2583353519439697, "reward_std": 0.34588029980659485, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8916687369346619, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7306250333786011, "step": 4592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 663.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 444.625, "completions/mean_terminated_length": 444.625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.9186, "grad_norm": 0.1314217448234558, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54391466.0, "reward": 1.3463540077209473, "reward_std": 0.34277278184890747, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8802062273025513, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 4593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 439.4375, "completions/mean_terminated_length": 439.4375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.9188, "grad_norm": 0.15424934029579163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54402289.0, "reward": 0.6641415357589722, "reward_std": 0.24715615808963776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2657812237739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 4594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 359.9375, "completions/mean_terminated_length": 359.9375, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.919, "grad_norm": 0.691320538520813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54411936.0, "reward": 0.9008321762084961, "reward_std": 0.47860631346702576, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42666250467300415, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 632.9375, "completions/mean_terminated_length": 632.9375, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.9192, "grad_norm": 0.17597374320030212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54426991.0, "reward": 0.971250593662262, "reward_std": 0.3346770405769348, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6299999952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.550000011920929, "step": 4596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 418.625, "completions/mean_terminated_length": 418.625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.9194, "grad_norm": 0.369622141122818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54437833.0, "reward": 1.258347749710083, "reward_std": 0.4232596158981323, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7666937708854675, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "step": 4597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 311.5625, "completions/mean_terminated_length": 311.5625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.9196, "grad_norm": 0.22036294639110565, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54446306.0, "reward": 0.6033540964126587, "reward_std": 0.23946958780288696, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14420625567436218, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 595.0625, "completions/mean_terminated_length": 566.4666748046875, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.9198, "grad_norm": 0.39139270782470703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54459923.0, "reward": 1.31193208694458, "reward_std": 0.4408542215824127, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8738625049591064, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 4599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 469.0, "completions/mean_terminated_length": 469.0, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.92, "grad_norm": 0.6124131083488464, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54471843.0, "reward": 1.0220820903778076, "reward_std": 0.4926415979862213, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5441625118255615, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 531.5, "completions/mean_terminated_length": 531.5, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.9202, "grad_norm": 0.03297099098563194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54484347.0, "reward": 0.5672666430473328, "reward_std": 0.05528552085161209, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13453124463558197, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "step": 4601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 603.375, "completions/mean_terminated_length": 575.3333740234375, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.9204, "grad_norm": 0.09381265193223953, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54498017.0, "reward": 0.5284383296966553, "reward_std": 0.14515063166618347, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11937499791383743, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.875, "step": 4602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 483.0625, "completions/mean_terminated_length": 483.0625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.9206, "grad_norm": 0.4374869763851166, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54510226.0, "reward": 0.8746882677078247, "reward_std": 0.3744455575942993, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.49937498569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8125, "step": 4603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 310.5, "completions/mean_terminated_length": 310.5, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.9208, "grad_norm": 0.3171488344669342, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54519194.0, "reward": 0.724450945854187, "reward_std": 0.3842809200286865, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2613999843597412, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 337.8125, "completions/mean_terminated_length": 337.8125, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.921, "grad_norm": 0.40677404403686523, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54528695.0, "reward": 1.1694539785385132, "reward_std": 0.45067495107650757, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7139062881469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 443.0, "completions/mean_terminated_length": 404.2666931152344, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.9212, "grad_norm": 0.2830323278903961, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54539527.0, "reward": 0.868110179901123, "reward_std": 0.5168718695640564, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42371875047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8793749809265137, "step": 4606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 685.5625, "completions/mean_terminated_length": 663.0000610351562, "completions/min_length": 450.0, "completions/min_terminated_length": 450.0, "epoch": 0.9214, "grad_norm": 0.2820824980735779, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54554560.0, "reward": 0.9916410446166992, "reward_std": 0.3217725157737732, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.8582812547683716, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.2593750059604645, "step": 4607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 689.75, "completions/mean_terminated_length": 667.4666748046875, "completions/min_length": 460.0, "completions/min_terminated_length": 460.0, "epoch": 0.9216, "grad_norm": 0.3950214087963104, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54569484.0, "reward": 0.8705477714538574, "reward_std": 0.5204991698265076, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4285937547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "step": 4608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 708.625, "completions/mean_terminated_length": 687.6000366210938, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.9218, "grad_norm": 0.2443838268518448, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54585926.0, "reward": 1.017319679260254, "reward_std": 0.5286498665809631, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5971375107765198, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8412500023841858, "step": 4609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 327.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 283.875, "completions/mean_terminated_length": 283.875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.922, "grad_norm": 0.466675341129303, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54594852.0, "reward": 0.7230228185653687, "reward_std": 0.3843725621700287, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2585437297821045, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 4610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 325.5625, "completions/mean_terminated_length": 325.5625, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.9222, "grad_norm": 0.0006119212484918535, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54604141.0, "reward": 1.4991728067398071, "reward_std": 0.0005645065102726221, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9983437061309814, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 750.1875, "completions/mean_terminated_length": 711.0714721679688, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.9224, "grad_norm": 0.33233770728111267, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54620144.0, "reward": 0.8471601009368896, "reward_std": 0.5262314677238464, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5068187713623047, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.715624988079071, "step": 4612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 863.5, "completions/mean_terminated_length": 826.4615478515625, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.9226, "grad_norm": 0.5738913416862488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54637976.0, "reward": 0.7967194318771362, "reward_std": 0.5643062591552734, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.468437522649765, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.65625, "step": 4613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 263.9375, "completions/mean_terminated_length": 263.9375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.9228, "grad_norm": 0.648639976978302, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54646327.0, "reward": 1.3777072429656982, "reward_std": 0.33124029636383057, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8804125189781189, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 382.875, "completions/mean_terminated_length": 382.875, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.923, "grad_norm": 0.01734856702387333, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54656501.0, "reward": 0.53000408411026, "reward_std": 0.017491210252046585, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06000624969601631, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 4615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 405.125, "completions/mean_terminated_length": 405.125, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.9232, "grad_norm": 0.09151770174503326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54667255.0, "reward": 0.5993759036064148, "reward_std": 0.08974257111549377, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19874998927116394, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 514.0, "completions/mean_terminated_length": 514.0, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.9234, "grad_norm": 0.4287852346897125, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54679463.0, "reward": 0.9654572010040283, "reward_std": 0.48594731092453003, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4934125244617462, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 515.6875, "completions/mean_terminated_length": 515.6875, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.9236, "grad_norm": 0.19792839884757996, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54691506.0, "reward": 1.4383478164672852, "reward_std": 0.23727308213710785, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9391937255859375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 942.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 572.1875, "completions/mean_terminated_length": 572.1875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.9238, "grad_norm": 0.2365836501121521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54704437.0, "reward": 0.7379693388938904, "reward_std": 0.2165328562259674, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47593751549720764, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6187499761581421, "step": 4619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 331.0625, "completions/mean_terminated_length": 331.0625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.924, "grad_norm": 0.016303859651088715, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54713590.0, "reward": 0.5226916074752808, "reward_std": 0.018306711688637733, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.045381247997283936, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750071525574, "step": 4620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 446.8125, "completions/mean_terminated_length": 446.8125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.9242, "grad_norm": 0.3966255187988281, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54724531.0, "reward": 0.889141321182251, "reward_std": 0.38422900438308716, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4657812714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 4621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 433.5, "completions/mean_terminated_length": 433.5, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.9244, "grad_norm": 0.3639155924320221, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54735211.0, "reward": 0.9644415378570557, "reward_std": 0.48680707812309265, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4913812577724457, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 405.875, "completions/mean_terminated_length": 405.875, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.9246, "grad_norm": 0.11079481989145279, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54746441.0, "reward": 0.5596104264259338, "reward_std": 0.0713874027132988, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11921874433755875, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 545.75, "completions/mean_terminated_length": 545.75, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.9248, "grad_norm": 0.3734476566314697, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54759269.0, "reward": 0.8477634787559509, "reward_std": 0.45429307222366333, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38302499055862427, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512499570846558, "step": 4624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 736.375, "completions/mean_terminated_length": 670.0, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.925, "grad_norm": 0.31574445962905884, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54775067.0, "reward": 0.505781888961792, "reward_std": 0.2890971302986145, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1990624964237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.668749988079071, "step": 4625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 448.75, "completions/mean_terminated_length": 448.75, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.9252, "grad_norm": 0.00046208815183490515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54786023.0, "reward": 1.4989041090011597, "reward_std": 0.0004187046142760664, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9978062510490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 300.375, "completions/mean_terminated_length": 300.375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.9254, "grad_norm": 0.22391407191753387, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54794557.0, "reward": 1.442798137664795, "reward_std": 0.22608007490634918, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9480937719345093, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "step": 4627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 384.0, "completions/max_terminated_length": 384.0, "completions/mean_length": 307.5625, "completions/mean_terminated_length": 307.5625, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.9256, "grad_norm": 0.00870053842663765, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54803046.0, "reward": 0.5469415187835693, "reward_std": 0.007237497717142105, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09388124942779541, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 334.0, "completions/max_terminated_length": 334.0, "completions/mean_length": 266.4375, "completions/mean_terminated_length": 266.4375, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.9258, "grad_norm": 0.0007370292441919446, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54812125.0, "reward": 1.4993884563446045, "reward_std": 0.0005835470510646701, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9987750053405762, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 4629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 303.125, "completions/mean_terminated_length": 303.125, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.926, "grad_norm": 0.6781876087188721, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54822175.0, "reward": 1.141772747039795, "reward_std": 0.47237515449523926, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6585437655448914, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9381250143051147, "step": 4630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 322.0, "completions/mean_terminated_length": 322.0, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.9262, "grad_norm": 0.24213428795337677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54831231.0, "reward": 1.4399665594100952, "reward_std": 0.23765823245048523, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9424312710762024, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "step": 4631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 262.0, "completions/max_terminated_length": 262.0, "completions/mean_length": 192.3125, "completions/mean_terminated_length": 192.3125, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 0.9264, "grad_norm": 1.214526653289795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54837956.0, "reward": 0.9811384081840515, "reward_std": 0.47599801421165466, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5247750282287598, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "step": 4632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 875.25, "completions/mean_terminated_length": 759.5555419921875, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.9266, "grad_norm": 0.7665743231773376, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54855944.0, "reward": 0.4640628695487976, "reward_std": 0.4813917875289917, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3031249940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.375, "step": 4633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 875.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 705.1875, "completions/mean_terminated_length": 705.1875, "completions/min_length": 607.0, "completions/min_terminated_length": 607.0, "epoch": 0.9268, "grad_norm": 0.2513248324394226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54871051.0, "reward": 1.0562384128570557, "reward_std": 0.46386709809303284, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6124750375747681, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 4634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 631.375, "completions/mean_terminated_length": 631.375, "completions/min_length": 521.0, "completions/min_terminated_length": 521.0, "epoch": 0.927, "grad_norm": 0.3772214949131012, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54884913.0, "reward": 1.1709258556365967, "reward_std": 0.4465271532535553, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7168500423431396, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 4635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 568.4375, "completions/mean_terminated_length": 568.4375, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.9272, "grad_norm": 0.47059497237205505, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54897576.0, "reward": 1.0444538593292236, "reward_std": 0.4405583441257477, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6514062881469727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7843749523162842, "step": 4636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 860.0, "completions/mean_length": 749.4375, "completions/mean_terminated_length": 731.1333618164062, "completions/min_length": 605.0, "completions/min_terminated_length": 605.0, "epoch": 0.9274, "grad_norm": 0.4099459648132324, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54913727.0, "reward": 1.1285853385925293, "reward_std": 0.5268463492393494, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6946687698364258, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8568750023841858, "step": 4637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 507.5, "completions/mean_terminated_length": 507.5, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.9276, "grad_norm": 0.28311190009117126, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54925991.0, "reward": 0.6638165712356567, "reward_std": 0.2702402174472809, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.26513123512268066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8456249833106995, "step": 4638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 654.9375, "completions/mean_terminated_length": 630.3333740234375, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.9278, "grad_norm": 0.15887218713760376, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54940310.0, "reward": 0.72816002368927, "reward_std": 0.30664607882499695, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.45631876587867737, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6731250286102295, "step": 4639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 856.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 551.4375, "completions/mean_terminated_length": 551.4375, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.928, "grad_norm": 0.09761693328619003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54952925.0, "reward": 0.6871757507324219, "reward_std": 0.19630347192287445, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.37435001134872437, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7049999237060547, "step": 4640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 647.8125, "completions/mean_terminated_length": 647.8125, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.9282, "grad_norm": 0.1810220330953598, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54967770.0, "reward": 0.9039853811264038, "reward_std": 0.4736519157886505, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43296873569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "step": 4641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 822.8125, "completions/mean_terminated_length": 794.0714721679688, "completions/min_length": 600.0, "completions/min_terminated_length": 600.0, "epoch": 0.9284, "grad_norm": 0.3536265194416046, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54985127.0, "reward": 1.0106914043426514, "reward_std": 0.5958104133605957, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5838812589645386, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.846875011920929, "step": 4642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 469.625, "completions/mean_terminated_length": 469.625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.9286, "grad_norm": 0.21626748144626617, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54996737.0, "reward": 0.814997673034668, "reward_std": 0.3683769702911377, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.442493736743927, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8243749737739563, "step": 4643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 679.125, "completions/mean_terminated_length": 599.5385131835938, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.9288, "grad_norm": 0.6976626515388489, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55011939.0, "reward": 1.0677945613861084, "reward_std": 0.6232152581214905, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6980875134468079, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7287499904632568, "step": 4644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 713.4375, "completions/mean_terminated_length": 669.0714721679688, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.929, "grad_norm": 0.36819252371788025, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55027050.0, "reward": 0.9525008201599121, "reward_std": 0.5521982312202454, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5924999713897705, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7625000476837158, "step": 4645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 841.6875, "completions/mean_terminated_length": 780.9166870117188, "completions/min_length": 513.0, "completions/min_terminated_length": 513.0, "epoch": 0.9292, "grad_norm": 0.3992539942264557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55044517.0, "reward": 0.8706130981445312, "reward_std": 0.6380218267440796, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.553725004196167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6487500071525574, "step": 4646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 431.1875, "completions/mean_terminated_length": 431.1875, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.9294, "grad_norm": 0.20404265820980072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55055480.0, "reward": 0.5948666334152222, "reward_std": 0.24198774993419647, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12723125517368317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "step": 4647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 963.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 628.4375, "completions/mean_terminated_length": 628.4375, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.9296, "grad_norm": 0.3919340968132019, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55069599.0, "reward": 1.2503008842468262, "reward_std": 0.3968789577484131, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8130999803543091, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8737499713897705, "step": 4648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 402.8125, "completions/mean_terminated_length": 402.8125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.9298, "grad_norm": 0.26604604721069336, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55079948.0, "reward": 0.7263287305831909, "reward_std": 0.205141082406044, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.45265626907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.640625, "step": 4649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 729.75, "completions/mean_terminated_length": 710.1333618164062, "completions/min_length": 573.0, "completions/min_terminated_length": 573.0, "epoch": 0.93, "grad_norm": 0.1478896290063858, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55095656.0, "reward": 0.8873163461685181, "reward_std": 0.4342435896396637, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5871312618255615, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6968749761581421, "step": 4650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 397.875, "completions/mean_terminated_length": 397.875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.9302, "grad_norm": 0.470071017742157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55106214.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 370.4375, "completions/mean_terminated_length": 370.4375, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.9304, "grad_norm": 0.4297231137752533, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55115981.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 788.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 539.125, "completions/mean_terminated_length": 539.125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.9306, "grad_norm": 0.19359943270683289, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55128287.0, "reward": 0.8756256103515625, "reward_std": 0.289015531539917, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5637499690055847, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6499999761581421, "step": 4653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.9308, "grad_norm": 0.43643954396247864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55138468.0, "reward": 0.7557603716850281, "reward_std": 0.382061630487442, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3240187466144562, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 4654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 415.0, "completions/max_terminated_length": 415.0, "completions/mean_length": 334.6875, "completions/mean_terminated_length": 334.6875, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.931, "grad_norm": 0.45785051584243774, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55148063.0, "reward": 1.3831665515899658, "reward_std": 0.3168083727359772, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8913312554359436, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 846.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 543.9375, "completions/mean_terminated_length": 543.9375, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.9312, "grad_norm": 0.10739853978157043, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55161054.0, "reward": 0.6443759202957153, "reward_std": 0.163855642080307, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2887499928474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7875000238418579, "step": 4656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 301.25, "completions/mean_terminated_length": 301.25, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.9314, "grad_norm": 0.515383243560791, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55170114.0, "reward": 1.3775384426116943, "reward_std": 0.3311760127544403, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8800750374794006, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 4657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 886.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 572.5, "completions/mean_terminated_length": 572.5, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.9316, "grad_norm": 0.21032772958278656, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55184138.0, "reward": 1.0346096754074097, "reward_std": 0.19474971294403076, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.8817187547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.24062500894069672, "step": 4658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 350.0, "completions/mean_terminated_length": 350.0, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.9318, "grad_norm": 0.19742032885551453, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55193786.0, "reward": 1.4653915166854858, "reward_std": 0.1324409544467926, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9932812452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 775.0, "completions/mean_terminated_length": 717.5385131835938, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.932, "grad_norm": 0.19932757318019867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55210938.0, "reward": 0.829687774181366, "reward_std": 0.3358849287033081, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7843749523162842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.21250000596046448, "step": 4660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 293.0, "completions/max_terminated_length": 293.0, "completions/mean_length": 248.625, "completions/mean_terminated_length": 248.625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.9322, "grad_norm": 0.42479002475738525, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55218372.0, "reward": 1.437232255935669, "reward_std": 0.24727551639080048, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9369624853134155, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 298.0, "completions/max_terminated_length": 298.0, "completions/mean_length": 243.75, "completions/mean_terminated_length": 243.75, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.9324, "grad_norm": 0.2769309878349304, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55226320.0, "reward": 1.440704107284546, "reward_std": 0.22552946209907532, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9439062476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281249642372131, "step": 4662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 602.0, "completions/mean_terminated_length": 602.0, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.9326, "grad_norm": 0.4725475013256073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55240000.0, "reward": 0.716547966003418, "reward_std": 0.38851019740104675, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24559375643730164, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 4663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 794.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 593.9375, "completions/mean_terminated_length": 593.9375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.9328, "grad_norm": 0.3873785734176636, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55253503.0, "reward": 0.9299225807189941, "reward_std": 0.36678677797317505, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6098437309265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.640625, "step": 4664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 557.875, "completions/mean_terminated_length": 526.800048828125, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.933, "grad_norm": 0.42353734374046326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55265981.0, "reward": 0.7012509107589722, "reward_std": 0.4197574853897095, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2775000035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8812499642372131, "step": 4665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 665.125, "completions/mean_terminated_length": 613.857177734375, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.9332, "grad_norm": 0.5504146814346313, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55280511.0, "reward": 0.8010040521621704, "reward_std": 0.5286357998847961, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4145062565803528, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8081250190734863, "step": 4666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1010.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 455.1875, "completions/mean_terminated_length": 455.1875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.9334, "grad_norm": 0.2140456736087799, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55292146.0, "reward": 1.1394509077072144, "reward_std": 0.44050368666648865, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7164000272750854, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8587499856948853, "step": 4667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 616.9375, "completions/mean_terminated_length": 616.9375, "completions/min_length": 475.0, "completions/min_terminated_length": 475.0, "epoch": 0.9336, "grad_norm": 0.43093428015708923, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55306049.0, "reward": 1.261447787284851, "reward_std": 0.42497050762176514, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7728937864303589, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 363.0, "completions/max_terminated_length": 363.0, "completions/mean_length": 301.875, "completions/mean_terminated_length": 301.875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.9338, "grad_norm": 0.8164277672767639, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55314927.0, "reward": 0.9610478281974792, "reward_std": 0.49045538902282715, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4845937490463257, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 481.5625, "completions/mean_terminated_length": 481.5625, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.934, "grad_norm": 0.013581285253167152, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55326440.0, "reward": 0.5396105647087097, "reward_std": 0.019651303067803383, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07921874523162842, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 4670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 758.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 457.1875, "completions/mean_terminated_length": 457.1875, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.9342, "grad_norm": 0.2167748212814331, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55338555.0, "reward": 0.6074227690696716, "reward_std": 0.12991376221179962, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21484375, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "step": 4671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 819.0, "completions/max_terminated_length": 819.0, "completions/mean_length": 546.25, "completions/mean_terminated_length": 546.25, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.9344, "grad_norm": 0.4010475277900696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55351503.0, "reward": 1.0408601760864258, "reward_std": 0.47656235098838806, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7067187428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.778124988079071, "step": 4672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 609.875, "completions/mean_terminated_length": 609.875, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.9346, "grad_norm": 0.1470920890569687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55365901.0, "reward": 0.8910941481590271, "reward_std": 0.17563816905021667, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7821874618530273, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3375000059604645, "step": 4673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 887.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 444.125, "completions/mean_terminated_length": 444.125, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.9348, "grad_norm": 0.28293558955192566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55377055.0, "reward": 1.3143510818481445, "reward_std": 0.3020491600036621, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.9412000179290771, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7537500262260437, "step": 4674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 692.25, "completions/mean_terminated_length": 670.1333618164062, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.935, "grad_norm": 0.5674748420715332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55392131.0, "reward": 0.6868758201599121, "reward_std": 0.3702511191368103, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3112500011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187499642372131, "step": 4675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 737.25, "completions/mean_terminated_length": 718.1333618164062, "completions/min_length": 423.0, "completions/min_terminated_length": 423.0, "epoch": 0.9352, "grad_norm": 0.1950404942035675, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55408295.0, "reward": 0.6260038614273071, "reward_std": 0.24632273614406586, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3145062327384949, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7081249952316284, "step": 4676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 512.0, "completions/mean_terminated_length": 477.86669921875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.9354, "grad_norm": 0.12872953712940216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55420455.0, "reward": 0.9367971420288086, "reward_std": 0.29670554399490356, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.8735937476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.18437500298023224, "step": 4677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 815.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 595.875, "completions/mean_terminated_length": 595.875, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.9356, "grad_norm": 0.20463870465755463, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55434229.0, "reward": 0.7877978086471558, "reward_std": 0.42444685101509094, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3255937695503235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 4678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 274.4375, "completions/mean_terminated_length": 274.4375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.9358, "grad_norm": 0.8607150912284851, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55442988.0, "reward": 0.6732697486877441, "reward_std": 0.32547762989997864, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2215375006198883, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "step": 4679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 452.5625, "completions/mean_terminated_length": 452.5625, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.936, "grad_norm": 0.23614242672920227, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55454117.0, "reward": 0.6178133487701416, "reward_std": 0.14230045676231384, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2356249988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 4680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 455.0, "completions/mean_terminated_length": 455.0, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.9362, "grad_norm": 0.306951642036438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55465989.0, "reward": 1.199497938156128, "reward_std": 0.4600489139556885, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.711493730545044, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "step": 4681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 579.5, "completions/mean_terminated_length": 549.86669921875, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.9364, "grad_norm": 0.34390366077423096, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55479261.0, "reward": 0.9642195105552673, "reward_std": 0.41202569007873535, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.7409375309944153, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.45625001192092896, "step": 4682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 811.9375, "completions/mean_terminated_length": 781.6428833007812, "completions/min_length": 620.0, "completions/min_terminated_length": 620.0, "epoch": 0.9366, "grad_norm": 0.37465700507164, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55495932.0, "reward": 0.6318758130073547, "reward_std": 0.4076775014400482, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.26374998688697815, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.78125, "step": 4683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 447.25, "completions/mean_terminated_length": 447.25, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.9368, "grad_norm": 0.319830060005188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55506752.0, "reward": 1.285775899887085, "reward_std": 0.39417099952697754, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8215500116348267, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 4684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 882.5625, "completions/mean_terminated_length": 818.2727661132812, "completions/min_length": 559.0, "completions/min_terminated_length": 559.0, "epoch": 0.937, "grad_norm": 0.530174970626831, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55525945.0, "reward": 0.6442190408706665, "reward_std": 0.4915889799594879, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5384374856948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.29374998807907104, "step": 4685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 561.25, "completions/mean_terminated_length": 530.4000244140625, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.9372, "grad_norm": 0.328363299369812, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55538589.0, "reward": 0.7785944938659668, "reward_std": 0.4066615104675293, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4321874976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7749999761581421, "step": 4686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 610.1875, "completions/mean_terminated_length": 514.6923217773438, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.9374, "grad_norm": 0.24179024994373322, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55552384.0, "reward": 0.7906255125999451, "reward_std": 0.4738408625125885, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.643750011920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.33125001192092896, "step": 4687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 310.0, "completions/max_terminated_length": 310.0, "completions/mean_length": 247.0, "completions/mean_terminated_length": 247.0, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.9376, "grad_norm": 0.00044901069486513734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55560752.0, "reward": 1.499100923538208, "reward_std": 0.0005386320990510285, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998199999332428, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "step": 4688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 415.5, "completions/mean_terminated_length": 415.5, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.9378, "grad_norm": 0.32068735361099243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55571096.0, "reward": 0.8142974376678467, "reward_std": 0.15802006423473358, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6285937428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5343749523162842, "step": 4689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 310.0, "completions/max_terminated_length": 310.0, "completions/mean_length": 246.75, "completions/mean_terminated_length": 246.75, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.938, "grad_norm": 0.0008091733907349408, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55578788.0, "reward": 1.499495029449463, "reward_std": 0.0012672666925936937, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989874362945557, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.981249988079071, "step": 4690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 555.0, "completions/max_terminated_length": 555.0, "completions/mean_length": 415.4375, "completions/mean_terminated_length": 415.4375, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.9382, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55589755.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 360.0, "completions/mean_terminated_length": 360.0, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.9384, "grad_norm": 0.46602317690849304, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55599707.0, "reward": 1.1225789785385132, "reward_std": 0.4656549096107483, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6201562881469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 646.125, "completions/mean_terminated_length": 646.125, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.9386, "grad_norm": 0.10872196406126022, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55614557.0, "reward": 1.0028910636901855, "reward_std": 0.22823180258274078, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8807812333106995, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3343749940395355, "step": 4693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 780.0, "completions/max_terminated_length": 780.0, "completions/mean_length": 611.75, "completions/mean_terminated_length": 611.75, "completions/min_length": 403.0, "completions/min_terminated_length": 403.0, "epoch": 0.9388, "grad_norm": 0.15060532093048096, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55628921.0, "reward": 0.6516698598861694, "reward_std": 0.3318360149860382, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17833749949932098, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "step": 4694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 718.25, "completions/mean_terminated_length": 718.25, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.939, "grad_norm": 0.14466369152069092, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55644445.0, "reward": 0.7557040452957153, "reward_std": 0.376649409532547, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3239062428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.909375011920929, "step": 4695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 725.125, "completions/mean_terminated_length": 682.4285888671875, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.9392, "grad_norm": 0.1173793226480484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55660095.0, "reward": 0.6220195293426514, "reward_std": 0.36752253770828247, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3065375089645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6612499952316284, "step": 4696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 323.3125, "completions/mean_terminated_length": 323.3125, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.9394, "grad_norm": 0.06309575587511063, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55670004.0, "reward": 0.9560944437980652, "reward_std": 0.08168401569128036, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9121875166893005, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.17500001192092896, "step": 4697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 726.875, "completions/mean_terminated_length": 726.875, "completions/min_length": 522.0, "completions/min_terminated_length": 522.0, "epoch": 0.9396, "grad_norm": 0.11266686767339706, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55686882.0, "reward": 0.814791202545166, "reward_std": 0.17552286386489868, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6295812129974365, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5081250071525574, "step": 4698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 517.0625, "completions/mean_terminated_length": 517.0625, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.9398, "grad_norm": 0.02029464580118656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55698883.0, "reward": 0.556079089641571, "reward_std": 0.02587219700217247, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11215624213218689, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "step": 4699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 797.0625, "completions/mean_terminated_length": 660.9000244140625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.94, "grad_norm": 0.8534792065620422, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55716740.0, "reward": 0.6850787401199341, "reward_std": 0.6223140954971313, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.3076562285423279, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.34687501192092896, "step": 4700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 431.6875, "completions/mean_terminated_length": 431.6875, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.9402, "grad_norm": 0.43909314274787903, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55727423.0, "reward": 1.099682092666626, "reward_std": 0.43890440464019775, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6993625164031982, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8299999833106995, "step": 4701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 947.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 455.5625, "completions/mean_terminated_length": 455.5625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.9404, "grad_norm": 0.35348236560821533, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55738472.0, "reward": 1.333150863647461, "reward_std": 0.30658218264579773, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.9162999987602234, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8475000262260437, "step": 4702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 463.625, "completions/mean_terminated_length": 463.625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.9406, "grad_norm": 0.08200615644454956, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55749922.0, "reward": 0.9967972040176392, "reward_std": 0.23663434386253357, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8685937523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.34687501192092896, "step": 4703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 521.5, "completions/mean_terminated_length": 521.5, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.9408, "grad_norm": 0.26070308685302734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55762538.0, "reward": 0.611016571521759, "reward_std": 0.23701012134552002, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15953123569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 587.25, "completions/mean_terminated_length": 558.1333618164062, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.941, "grad_norm": 0.19228820502758026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55776798.0, "reward": 0.7212506532669067, "reward_std": 0.3214277923107147, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4424999952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6625000238418579, "step": 4705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 843.9375, "completions/mean_terminated_length": 762.0909423828125, "completions/min_length": 504.0, "completions/min_terminated_length": 504.0, "epoch": 0.9412, "grad_norm": 0.2548653483390808, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55794717.0, "reward": 0.5715630650520325, "reward_std": 0.5990365743637085, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3931250274181366, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4437500238418579, "step": 4706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 384.1875, "completions/mean_terminated_length": 384.1875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.9414, "grad_norm": 0.10816261172294617, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55804976.0, "reward": 0.5838571190834045, "reward_std": 0.1217232421040535, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1677124947309494, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8949999809265137, "step": 4707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 308.125, "completions/mean_terminated_length": 308.125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.9416, "grad_norm": 0.0007149228476919234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55813922.0, "reward": 1.4989041090011597, "reward_std": 0.0004187046142760664, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9978062510490417, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 328.1875, "completions/mean_terminated_length": 328.1875, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.9418, "grad_norm": 0.09214316308498383, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55823413.0, "reward": 0.5876540541648865, "reward_std": 0.05889489874243736, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17530623078346252, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "step": 4709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 557.0, "completions/mean_terminated_length": 557.0, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.942, "grad_norm": 0.17025531828403473, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55836197.0, "reward": 0.9084227681159973, "reward_std": 0.4727901518344879, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44184374809265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9481250047683716, "step": 4710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 372.0, "completions/max_terminated_length": 372.0, "completions/mean_length": 273.25, "completions/mean_terminated_length": 273.25, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.9422, "grad_norm": 0.0005713177379220724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55845001.0, "reward": 1.4992196559906006, "reward_std": 0.0005501238629221916, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9984375238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 675.0, "completions/max_terminated_length": 675.0, "completions/mean_length": 568.5625, "completions/mean_terminated_length": 568.5625, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.9424, "grad_norm": 0.22956135869026184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55857938.0, "reward": 1.3802634477615356, "reward_std": 0.3245895504951477, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8855249881744385, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "step": 4712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 352.1875, "completions/mean_terminated_length": 352.1875, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.9426, "grad_norm": 0.436300665140152, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55867493.0, "reward": 1.044219732284546, "reward_std": 0.47616440057754517, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5884374976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 4713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 307.0, "completions/max_terminated_length": 307.0, "completions/mean_length": 274.25, "completions/mean_terminated_length": 274.25, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.9428, "grad_norm": 0.0004227568279020488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55876073.0, "reward": 1.4990415573120117, "reward_std": 0.0005209985538385808, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980812668800354, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 634.5, "completions/mean_terminated_length": 578.857177734375, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.943, "grad_norm": 0.14673182368278503, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55890113.0, "reward": 0.6640630960464478, "reward_std": 0.36837247014045715, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.390625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.625, "step": 4715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 410.75, "completions/mean_terminated_length": 410.75, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.9432, "grad_norm": 0.19089648127555847, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55900221.0, "reward": 1.4342821836471558, "reward_std": 0.24673327803611755, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9310625195503235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 352.0, "completions/max_terminated_length": 352.0, "completions/mean_length": 263.5, "completions/mean_terminated_length": 263.5, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.9434, "grad_norm": 0.013093648478388786, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55908053.0, "reward": 0.5480228662490845, "reward_std": 0.015930648893117905, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09604375064373016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "step": 4717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 374.3125, "completions/mean_terminated_length": 374.3125, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.9436, "grad_norm": 0.36162030696868896, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55918266.0, "reward": 0.9137507677078247, "reward_std": 0.4302397072315216, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5149999856948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "step": 4718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 352.0, "completions/mean_terminated_length": 352.0, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.9438, "grad_norm": 0.6070457100868225, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55927450.0, "reward": 1.2570104598999023, "reward_std": 0.43400824069976807, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.764018714427948, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750667572021, "step": 4719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 404.0, "completions/max_terminated_length": 404.0, "completions/mean_length": 302.3125, "completions/mean_terminated_length": 302.3125, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.944, "grad_norm": 0.2772347629070282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55936031.0, "reward": 0.6913164258003235, "reward_std": 0.24509860575199127, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.32013124227523804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7893749475479126, "step": 4720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 849.0, "completions/max_terminated_length": 849.0, "completions/mean_length": 633.625, "completions/mean_terminated_length": 633.625, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.9442, "grad_norm": 0.18272104859352112, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55950425.0, "reward": 1.3061978816986084, "reward_std": 0.3556644916534424, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8623937368392944, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8818750381469727, "step": 4721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 326.0, "completions/mean_terminated_length": 326.0, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.9444, "grad_norm": 0.26124465465545654, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55959529.0, "reward": 1.3803728818893433, "reward_std": 0.324632465839386, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8857437372207642, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 373.5625, "completions/mean_terminated_length": 373.5625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.9446, "grad_norm": 0.012513219378888607, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55969090.0, "reward": 0.5457041263580322, "reward_std": 0.012187209911644459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09140624850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "step": 4723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 383.4375, "completions/mean_terminated_length": 383.4375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.9448, "grad_norm": 0.13857682049274445, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55979321.0, "reward": 1.442266583442688, "reward_std": 0.22593748569488525, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9470312595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 668.75, "completions/mean_terminated_length": 645.0667114257812, "completions/min_length": 513.0, "completions/min_terminated_length": 513.0, "epoch": 0.945, "grad_norm": 0.23137541115283966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55994133.0, "reward": 1.0181382894515991, "reward_std": 0.5296979546546936, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5987750291824341, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.846250057220459, "step": 4725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 644.3125, "completions/mean_terminated_length": 619.0000610351562, "completions/min_length": 432.0, "completions/min_terminated_length": 432.0, "epoch": 0.9452, "grad_norm": 0.06618838012218475, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56009210.0, "reward": 0.613906979560852, "reward_std": 0.23755642771720886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2903124988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7249999642372131, "step": 4726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 664.9375, "completions/mean_terminated_length": 641.0000610351562, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.9454, "grad_norm": 0.1715490221977234, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56023625.0, "reward": 1.227097749710083, "reward_std": 0.501533031463623, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7666937708854675, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8956249952316284, "step": 4727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 456.5, "completions/mean_terminated_length": 375.4285888671875, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.9456, "grad_norm": 0.33691948652267456, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56034545.0, "reward": 0.5786726474761963, "reward_std": 0.3268779516220093, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21984374523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7843749523162842, "step": 4728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 407.0625, "completions/mean_terminated_length": 407.0625, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.9458, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56044994.0, "reward": 1.498750925064087, "reward_std": 0.0, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9975000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 436.0, "completions/max_terminated_length": 436.0, "completions/mean_length": 352.9375, "completions/mean_terminated_length": 352.9375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.946, "grad_norm": 0.2957916259765625, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56054513.0, "reward": 0.8325603008270264, "reward_std": 0.46453049778938293, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3526187539100647, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706250429153442, "step": 4730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 600.8125, "completions/mean_terminated_length": 572.6000366210938, "completions/min_length": 423.0, "completions/min_terminated_length": 423.0, "epoch": 0.9462, "grad_norm": 0.17020076513290405, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56068286.0, "reward": 1.2325539588928223, "reward_std": 0.4852559268474579, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7776062488555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.856874942779541, "step": 4731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 573.125, "completions/mean_terminated_length": 508.71429443359375, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.9464, "grad_norm": 0.2944878935813904, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56081088.0, "reward": 1.1442071199417114, "reward_std": 0.5698489546775818, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7259125113487244, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8174999952316284, "step": 4732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 994.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 619.5, "completions/mean_terminated_length": 619.5, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.9466, "grad_norm": 0.12371402233839035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56094600.0, "reward": 0.7946101427078247, "reward_std": 0.323403537273407, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.46421873569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.765625, "step": 4733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 629.5, "completions/mean_terminated_length": 603.2000122070312, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.9468, "grad_norm": 0.24122542142868042, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56108976.0, "reward": 0.7136727571487427, "reward_std": 0.4189833402633667, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3023437559604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8656249642372131, "step": 4734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 567.8125, "completions/mean_terminated_length": 502.64288330078125, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.947, "grad_norm": 0.10509012639522552, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56122365.0, "reward": 0.5738288164138794, "reward_std": 0.280038982629776, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.27265626192092896, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6656249761581421, "step": 4735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 604.0, "completions/max_terminated_length": 604.0, "completions/mean_length": 355.8125, "completions/mean_terminated_length": 355.8125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.9472, "grad_norm": 0.4905503988265991, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56132138.0, "reward": 0.8720321655273438, "reward_std": 0.43894943594932556, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.43156251311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 4736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 771.5625, "completions/mean_terminated_length": 735.5000610351562, "completions/min_length": 469.0, "completions/min_terminated_length": 469.0, "epoch": 0.9474, "grad_norm": 0.1230834424495697, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56148611.0, "reward": 0.5085914731025696, "reward_std": 0.21875561773777008, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14218124747276306, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7931250333786011, "step": 4737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 529.625, "completions/mean_terminated_length": 529.625, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.9476, "grad_norm": 0.16931121051311493, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56162461.0, "reward": 0.9920977354049683, "reward_std": 0.430370569229126, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.609193742275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7893749475479126, "step": 4738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 468.625, "completions/mean_terminated_length": 468.625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.9478, "grad_norm": 0.23683170974254608, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56174647.0, "reward": 0.8485165238380432, "reward_std": 0.4529082477092743, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38453125953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 650.8125, "completions/mean_terminated_length": 597.5, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.948, "grad_norm": 0.37123939394950867, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56188708.0, "reward": 0.8835948705673218, "reward_std": 0.5851116180419922, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4546875059604645, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.84375, "step": 4740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 377.0, "completions/mean_terminated_length": 377.0, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.9482, "grad_norm": 0.21613119542598724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56199476.0, "reward": 0.6112477779388428, "reward_std": 0.2364569902420044, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1599937379360199, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "step": 4741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 453.125, "completions/mean_terminated_length": 453.125, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.9484, "grad_norm": 0.24871626496315002, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56210246.0, "reward": 0.7598665952682495, "reward_std": 0.44138291478157043, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26973122358322144, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9868749976158142, "step": 4742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 467.0625, "completions/mean_terminated_length": 467.0625, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.9486, "grad_norm": 0.0018336634384468198, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56221687.0, "reward": 1.4974102973937988, "reward_std": 0.004746776539832354, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9948187470436096, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9393750429153442, "step": 4743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 333.0625, "completions/mean_terminated_length": 333.0625, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.9488, "grad_norm": 0.3113432228565216, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56230808.0, "reward": 0.6038915514945984, "reward_std": 0.2389136105775833, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14528125524520874, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 4744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1022.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 582.25, "completions/mean_terminated_length": 582.25, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.949, "grad_norm": 0.17182444036006927, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56244124.0, "reward": 0.7917974591255188, "reward_std": 0.2108914852142334, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5835937261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5281250476837158, "step": 4745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 779.3125, "completions/mean_terminated_length": 722.84619140625, "completions/min_length": 565.0, "completions/min_terminated_length": 565.0, "epoch": 0.9492, "grad_norm": 0.22981826961040497, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56261153.0, "reward": 0.7421879768371582, "reward_std": 0.43637192249298096, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.609375, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.3125, "step": 4746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 436.3125, "completions/mean_terminated_length": 436.3125, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.9494, "grad_norm": 0.2839629650115967, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56272326.0, "reward": 1.436563491821289, "reward_std": 0.23700281977653503, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9356249570846558, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 4747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 742.125, "completions/mean_terminated_length": 701.857177734375, "completions/min_length": 522.0, "completions/min_terminated_length": 522.0, "epoch": 0.9496, "grad_norm": 0.23085308074951172, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56287816.0, "reward": 0.6297663450241089, "reward_std": 0.3429029583930969, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3220312297344208, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 4748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 434.5625, "completions/mean_terminated_length": 434.5625, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.9498, "grad_norm": 0.40749725699424744, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56298625.0, "reward": 0.8700790405273438, "reward_std": 0.4405728876590729, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.42765626311302185, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "step": 4749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 342.375, "completions/mean_terminated_length": 342.375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.95, "grad_norm": 0.17223061621189117, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56307735.0, "reward": 0.5763289928436279, "reward_std": 0.07759187370538712, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15265624225139618, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "step": 4750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 777.625, "completions/mean_terminated_length": 742.4285888671875, "completions/min_length": 584.0, "completions/min_terminated_length": 584.0, "epoch": 0.9502, "grad_norm": 0.25709930062294006, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56326225.0, "reward": 0.847500205039978, "reward_std": 0.3320743441581726, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8199999928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.20000001788139343, "step": 4751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 321.625, "completions/mean_terminated_length": 321.625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.9504, "grad_norm": 0.4933495819568634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56335179.0, "reward": 1.077731966972351, "reward_std": 0.3646717667579651, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.5304625034332275, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5712500214576721, "step": 4752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 583.25, "completions/mean_terminated_length": 583.25, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.9506, "grad_norm": 0.0330837108194828, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56348703.0, "reward": 0.5992196202278137, "reward_std": 0.0912434533238411, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19843749701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.887499988079071, "step": 4753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 298.75, "completions/mean_terminated_length": 298.75, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.9508, "grad_norm": 0.00020228829816915095, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56357515.0, "reward": 1.4988259077072144, "reward_std": 0.00030002594576217234, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976500272750854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 774.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 598.5625, "completions/mean_terminated_length": 598.5625, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.951, "grad_norm": 0.1879192739725113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56371796.0, "reward": 0.7771883010864258, "reward_std": 0.34073367714881897, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4293749928474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7374999523162842, "step": 4755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 482.5, "completions/mean_terminated_length": 482.5, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.9512, "grad_norm": 0.36248815059661865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56383484.0, "reward": 0.9982821345329285, "reward_std": 0.46344539523124695, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5590624809265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 4756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 620.0, "completions/max_terminated_length": 620.0, "completions/mean_length": 541.5625, "completions/mean_terminated_length": 541.5625, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.9514, "grad_norm": 0.15284088253974915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56396037.0, "reward": 1.3207197189331055, "reward_std": 0.3830054700374603, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8289375305175781, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 4757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 482.25, "completions/mean_terminated_length": 482.25, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.9516, "grad_norm": 0.2940591275691986, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56408521.0, "reward": 0.9643759727478027, "reward_std": 0.48673015832901, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49125000834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 435.6875, "completions/mean_terminated_length": 435.6875, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.9518, "grad_norm": 0.09862081706523895, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56419476.0, "reward": 0.6075009107589722, "reward_std": 0.23499998450279236, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1525000035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 365.25, "completions/mean_terminated_length": 365.25, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.952, "grad_norm": 0.09259848296642303, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56429192.0, "reward": 1.4906134605407715, "reward_std": 0.03085871785879135, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9812250137329102, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 4760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 385.0, "completions/max_terminated_length": 385.0, "completions/mean_length": 316.25, "completions/mean_terminated_length": 316.25, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.9522, "grad_norm": 0.19645877182483673, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56438204.0, "reward": 1.380157232284546, "reward_std": 0.31563717126846313, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8853124976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9124999642372131, "step": 4761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 944.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 611.4375, "completions/mean_terminated_length": 611.4375, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.9524, "grad_norm": 0.05081353709101677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56452467.0, "reward": 0.5824227333068848, "reward_std": 0.11113828420639038, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16484373807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 585.25, "completions/mean_terminated_length": 522.5714721679688, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.9526, "grad_norm": 0.15453729033470154, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56465895.0, "reward": 0.6828913688659668, "reward_std": 0.4259311258792877, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3657812476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6656249761581421, "step": 4763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 631.875, "completions/mean_terminated_length": 605.7333374023438, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.9528, "grad_norm": 0.11729738116264343, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56480629.0, "reward": 0.7089849710464478, "reward_std": 0.2568126320838928, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.48046875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5968749523162842, "step": 4764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 403.0, "completions/max_terminated_length": 403.0, "completions/mean_length": 314.625, "completions/mean_terminated_length": 314.625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.953, "grad_norm": 0.2544809877872467, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56489359.0, "reward": 1.3209104537963867, "reward_std": 0.38309866189956665, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8293187618255615, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 265.0, "completions/max_terminated_length": 265.0, "completions/mean_length": 178.6875, "completions/mean_terminated_length": 178.6875, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 0.9532, "grad_norm": 0.0013406979851424694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56496010.0, "reward": 1.499269723892212, "reward_std": 0.0012723120162263513, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.998537540435791, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "step": 4766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 477.0, "completions/mean_terminated_length": 477.0, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.9534, "grad_norm": 0.05515504628419876, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56507514.0, "reward": 0.9155471920967102, "reward_std": 0.09571963548660278, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8310937881469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3531250059604645, "step": 4767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 397.1875, "completions/mean_terminated_length": 397.1875, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.9536, "grad_norm": 0.0002511032216716558, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56517805.0, "reward": 1.4988821744918823, "reward_std": 0.0003586417296901345, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9977625012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 585.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 424.5, "completions/mean_terminated_length": 424.5, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.9538, "grad_norm": 0.27418455481529236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56528165.0, "reward": 1.0603070259094238, "reward_std": 0.33648601174354553, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.8081125020980835, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5862500071525574, "step": 4769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 921.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 526.6875, "completions/mean_terminated_length": 526.6875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.954, "grad_norm": 0.00264346064068377, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56540800.0, "reward": 0.9925001859664917, "reward_std": 0.006324579007923603, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9850000143051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.11249999701976776, "step": 4770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 585.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 281.6875, "completions/mean_terminated_length": 281.6875, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.9542, "grad_norm": 0.38855618238449097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56549307.0, "reward": 1.372260332107544, "reward_std": 0.27220141887664795, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9320187568664551, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8331249952316284, "step": 4771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 373.6875, "completions/mean_terminated_length": 373.6875, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.9544, "grad_norm": 0.437322199344635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56559062.0, "reward": 0.7862509489059448, "reward_std": 0.4248529076576233, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32249999046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 421.5, "completions/mean_terminated_length": 421.5, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.9546, "grad_norm": 0.03548271954059601, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56569998.0, "reward": 1.4816728830337524, "reward_std": 0.07011494785547256, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9633437395095825, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9206250309944153, "step": 4773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 275.0, "completions/mean_terminated_length": 275.0, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.9548, "grad_norm": 0.5891707539558411, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56578270.0, "reward": 0.8094447255134583, "reward_std": 0.42656996846199036, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3688875138759613, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9075000286102295, "step": 4774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 224.75, "completions/mean_terminated_length": 224.75, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.955, "grad_norm": 0.4043081998825073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56585706.0, "reward": 1.385469675064087, "reward_std": 0.3021526634693146, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8959375023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9249999523162842, "step": 4775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 879.5, "completions/mean_terminated_length": 813.8181762695312, "completions/min_length": 567.0, "completions/min_terminated_length": 567.0, "epoch": 0.9552, "grad_norm": 0.2807249426841736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56604018.0, "reward": 0.649375319480896, "reward_std": 0.45512980222702026, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6112500429153442, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.17500001192092896, "step": 4776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 451.625, "completions/mean_terminated_length": 451.625, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.9554, "grad_norm": 0.3026505410671234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56617356.0, "reward": 1.2000665664672852, "reward_std": 0.457598477602005, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7126312851905823, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 851.0, "completions/max_terminated_length": 851.0, "completions/mean_length": 444.3125, "completions/mean_terminated_length": 444.3125, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.9556, "grad_norm": 0.2294689267873764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56628577.0, "reward": 0.9268759489059448, "reward_std": 0.45550188422203064, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47874999046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8624999523162842, "step": 4778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 232.8125, "completions/mean_terminated_length": 232.8125, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.9558, "grad_norm": 0.575843095779419, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56635758.0, "reward": 0.7401381731033325, "reward_std": 0.27796241641044617, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5427749156951904, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5524374842643738, "step": 4779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1003.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 683.4375, "completions/mean_terminated_length": 683.4375, "completions/min_length": 459.0, "completions/min_terminated_length": 459.0, "epoch": 0.956, "grad_norm": 0.12421514093875885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56650741.0, "reward": 0.6971883773803711, "reward_std": 0.17327773571014404, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3943749964237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.737500011920929, "step": 4780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 285.0, "completions/mean_terminated_length": 285.0, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.9562, "grad_norm": 0.6822489500045776, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56658821.0, "reward": 1.1408321857452393, "reward_std": 0.4775739014148712, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6566624641418457, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 408.0, "completions/mean_terminated_length": 408.0, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.9564, "grad_norm": 0.4712301194667816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56669413.0, "reward": 1.0237510204315186, "reward_std": 0.49057790637016296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 442.0, "completions/mean_length": 377.5625, "completions/mean_terminated_length": 334.4666748046875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.9566, "grad_norm": 0.5481482744216919, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56679870.0, "reward": 1.16170072555542, "reward_std": 0.4979201555252075, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7609000205993652, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7912499904632568, "step": 4783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 544.375, "completions/mean_terminated_length": 544.375, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.9568, "grad_norm": 0.17555655539035797, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56692276.0, "reward": 1.0177572965621948, "reward_std": 0.49762770533561707, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5355125069618225, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "step": 4784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 308.25, "completions/mean_terminated_length": 308.25, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.957, "grad_norm": 0.6596409678459167, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56700936.0, "reward": 1.2173289060592651, "reward_std": 0.3407399654388428, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.8721562623977661, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6481249928474426, "step": 4785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 718.0, "completions/max_terminated_length": 718.0, "completions/mean_length": 451.625, "completions/mean_terminated_length": 451.625, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.9572, "grad_norm": 0.1567268967628479, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56711970.0, "reward": 0.9917190670967102, "reward_std": 0.19056597352027893, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8584375381469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3125, "step": 4786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 587.5, "completions/mean_terminated_length": 486.7692565917969, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.9574, "grad_norm": 0.3264961838722229, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56726714.0, "reward": 0.7722662091255188, "reward_std": 0.5277140140533447, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4820312261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6343749761581421, "step": 4787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 483.8125, "completions/mean_terminated_length": 483.8125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.9576, "grad_norm": 0.2685251533985138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56738663.0, "reward": 0.9613293409347534, "reward_std": 0.49027979373931885, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48515623807907104, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "step": 4788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 454.5, "completions/mean_terminated_length": 454.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.9578, "grad_norm": 0.07807733118534088, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56750175.0, "reward": 1.4599071741104126, "reward_std": 0.1559755653142929, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9823125004768372, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262499809265137, "step": 4789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 802.1875, "completions/mean_terminated_length": 770.5000610351562, "completions/min_length": 592.0, "completions/min_terminated_length": 592.0, "epoch": 0.958, "grad_norm": 0.2055627703666687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56767922.0, "reward": 0.5440506935119629, "reward_std": 0.2291020154953003, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21310000121593475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7549999952316284, "step": 4790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 384.0, "completions/mean_terminated_length": 384.0, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.9582, "grad_norm": 0.2473229169845581, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56778178.0, "reward": 0.7250665426254272, "reward_std": 0.38392210006713867, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26263123750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 712.5625, "completions/mean_terminated_length": 691.800048828125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.9584, "grad_norm": 0.1083318293094635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56793643.0, "reward": 0.6025789380073547, "reward_std": 0.21534161269664764, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20515623688697815, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7531249523162842, "step": 4792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 310.875, "completions/mean_terminated_length": 310.875, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.9586, "grad_norm": 0.22482794523239136, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56802409.0, "reward": 1.3800666332244873, "reward_std": 0.3245124816894531, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8851312398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 568.625, "completions/mean_terminated_length": 568.625, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.9588, "grad_norm": 0.24247439205646515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56816307.0, "reward": 1.0167973041534424, "reward_std": 0.24224121868610382, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.7835937738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.37187498807907104, "step": 4794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 479.875, "completions/mean_terminated_length": 479.875, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.959, "grad_norm": 0.4551601707935333, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56827537.0, "reward": 1.1299445629119873, "reward_std": 0.4498065412044525, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6973875164985657, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "step": 4795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 661.8125, "completions/mean_terminated_length": 578.2307739257812, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.9592, "grad_norm": 0.38504642248153687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56842318.0, "reward": 0.7698442935943604, "reward_std": 0.4968777298927307, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5396875143051147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5249999761581421, "step": 4796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 634.125, "completions/mean_terminated_length": 578.4285888671875, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.9594, "grad_norm": 0.5817534923553467, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56856112.0, "reward": 1.0205156803131104, "reward_std": 0.5824776887893677, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.603529691696167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8003125190734863, "step": 4797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 399.875, "completions/mean_terminated_length": 399.875, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.9596, "grad_norm": 0.3638065457344055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56866190.0, "reward": 0.9614853858947754, "reward_std": 0.4840942621231079, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4854687452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 969.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 414.25, "completions/mean_terminated_length": 414.25, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.9598, "grad_norm": 0.22703054547309875, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56876738.0, "reward": 0.7365633249282837, "reward_std": 0.2557372450828552, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.41062498092651367, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "step": 4799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 180.0625, "completions/mean_terminated_length": 180.0625, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.96, "grad_norm": 0.294859379529953, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56883235.0, "reward": 1.0743756294250488, "reward_std": 0.22082316875457764, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7737500071525574, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.26249998807907104, "step": 4800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 503.625, "completions/mean_terminated_length": 468.933349609375, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.9602, "grad_norm": 0.5605654716491699, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56895517.0, "reward": 0.7577351331710815, "reward_std": 0.4194486737251282, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3904687464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906249761581421, "step": 4801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 850.375, "completions/mean_terminated_length": 810.3077392578125, "completions/min_length": 566.0, "completions/min_terminated_length": 566.0, "epoch": 0.9604, "grad_norm": 0.24904386699199677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56914163.0, "reward": 0.5700787305831909, "reward_std": 0.3259633481502533, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32765623927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.559374988079071, "step": 4802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 909.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 511.375, "completions/mean_terminated_length": 511.375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.9606, "grad_norm": 0.12403276562690735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56925961.0, "reward": 1.3771603107452393, "reward_std": 0.3330119848251343, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8793187141418457, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 442.25, "completions/mean_terminated_length": 442.25, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.9608, "grad_norm": 0.14137612283229828, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56936637.0, "reward": 0.6101570725440979, "reward_std": 0.15188705921173096, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22031250596046448, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8374999761581421, "step": 4804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 804.875, "completions/mean_terminated_length": 754.3077392578125, "completions/min_length": 522.0, "completions/min_terminated_length": 522.0, "epoch": 0.961, "grad_norm": 0.18964186310768127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56953979.0, "reward": 0.5789850950241089, "reward_std": 0.3846968412399292, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.28296875953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.640625, "step": 4805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 543.0625, "completions/mean_terminated_length": 543.0625, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.9612, "grad_norm": 0.011878385208547115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56966588.0, "reward": 0.5353761911392212, "reward_std": 0.02099323645234108, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07074999809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "step": 4806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 317.8125, "completions/mean_terminated_length": 317.8125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.9614, "grad_norm": 0.0007321523153223097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56975513.0, "reward": 1.4990415573120117, "reward_std": 0.0005210168310441077, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980812668800354, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 454.0625, "completions/mean_terminated_length": 454.0625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.9616, "grad_norm": 0.3259725570678711, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56986938.0, "reward": 1.36570405960083, "reward_std": 0.3007901906967163, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9189062118530273, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.871874988079071, "step": 4808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 640.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 514.0, "completions/mean_terminated_length": 514.0, "completions/min_length": 424.0, "completions/min_terminated_length": 424.0, "epoch": 0.9618, "grad_norm": 0.47658270597457886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56999114.0, "reward": 0.9050009846687317, "reward_std": 0.4749999940395355, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4350000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 454.3125, "completions/mean_terminated_length": 454.3125, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.962, "grad_norm": 0.17108142375946045, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57010191.0, "reward": 1.1357038021087646, "reward_std": 0.29430529475212097, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.8964062333106995, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4593750238418579, "step": 4810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 380.1875, "completions/mean_terminated_length": 337.2666931152344, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.9622, "grad_norm": 0.2619990408420563, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57020706.0, "reward": 0.9896883964538574, "reward_std": 0.5375016331672668, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5418750047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.875, "step": 4811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 367.0, "completions/max_terminated_length": 367.0, "completions/mean_length": 333.875, "completions/mean_terminated_length": 333.875, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.9624, "grad_norm": 0.00034886185312643647, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57029936.0, "reward": 1.4988164901733398, "reward_std": 0.0002624988555908203, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9976312518119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 540.0, "completions/mean_terminated_length": 540.0, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.9626, "grad_norm": 0.321535587310791, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57042800.0, "reward": 0.6482696533203125, "reward_std": 0.24498561024665833, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.23403750360012054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 4813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 308.3125, "completions/mean_terminated_length": 308.3125, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.9628, "grad_norm": 0.17987936735153198, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57051749.0, "reward": 1.4424103498458862, "reward_std": 0.22597619891166687, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9473187923431396, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 742.25, "completions/mean_terminated_length": 723.4666748046875, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.963, "grad_norm": 0.4087540805339813, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57068297.0, "reward": 0.793203592300415, "reward_std": 0.2989886403083801, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6489062309265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.359375, "step": 4815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 452.125, "completions/mean_terminated_length": 414.0000305175781, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.9632, "grad_norm": 0.37912774085998535, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57079435.0, "reward": 0.8242946267127991, "reward_std": 0.4928596317768097, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39858752489089966, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8774999976158142, "step": 4816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 461.6875, "completions/mean_terminated_length": 461.6875, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.9634, "grad_norm": 0.5764101147651672, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57090710.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 316.0, "completions/max_terminated_length": 316.0, "completions/mean_length": 284.75, "completions/mean_terminated_length": 284.75, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.9636, "grad_norm": 0.20711080729961395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57098546.0, "reward": 1.3778572082519531, "reward_std": 0.3312990665435791, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8807125091552734, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 442.6875, "completions/mean_terminated_length": 442.6875, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.9638, "grad_norm": 0.3614601194858551, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57109757.0, "reward": 1.3800009489059448, "reward_std": 0.32448676228523254, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8849999904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 419.5625, "completions/mean_terminated_length": 419.5625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.964, "grad_norm": 0.3600819706916809, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57121350.0, "reward": 0.9096100330352783, "reward_std": 0.33919060230255127, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6317187547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.653124988079071, "step": 4820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 376.25, "completions/mean_terminated_length": 376.25, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.9642, "grad_norm": 0.3958362340927124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57131482.0, "reward": 1.0812883377075195, "reward_std": 0.4888736605644226, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6000750064849854, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 4821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 602.75, "completions/mean_terminated_length": 602.75, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 0.9644, "grad_norm": 0.2712974548339844, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57145894.0, "reward": 0.6782820224761963, "reward_std": 0.2682052254676819, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4190624952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6375000476837158, "step": 4822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 457.625, "completions/mean_terminated_length": 419.86669921875, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.9646, "grad_norm": 0.16677208244800568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57157744.0, "reward": 0.6323444843292236, "reward_std": 0.24882423877716064, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.32718750834465027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 4823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 490.5, "completions/mean_terminated_length": 454.933349609375, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.9648, "grad_norm": 0.49589377641677856, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57170024.0, "reward": 0.9299885034561157, "reward_std": 0.5353918671607971, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48497501015663147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8925000429153442, "step": 4824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 490.8125, "completions/mean_terminated_length": 490.8125, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.965, "grad_norm": 0.20713308453559875, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57182149.0, "reward": 0.843882143497467, "reward_std": 0.4561940133571625, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3752624988555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 975.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 508.0625, "completions/mean_terminated_length": 508.0625, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.9652, "grad_norm": 0.21685805916786194, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57194262.0, "reward": 0.8103131055831909, "reward_std": 0.1737101674079895, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6206250190734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.512499988079071, "step": 4826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 291.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 240.5, "completions/mean_terminated_length": 240.5, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.9654, "grad_norm": 0.34989434480667114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57202206.0, "reward": 0.9020323157310486, "reward_std": 0.47769105434417725, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42906248569488525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "step": 4827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 293.3125, "completions/mean_terminated_length": 293.3125, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.9656, "grad_norm": 0.18857769668102264, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57211059.0, "reward": 0.6081259250640869, "reward_std": 0.23749999701976776, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1537500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 673.625, "completions/mean_terminated_length": 623.5714721679688, "completions/min_length": 473.0, "completions/min_terminated_length": 473.0, "epoch": 0.9658, "grad_norm": 0.5207056999206543, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57225517.0, "reward": 0.8696101903915405, "reward_std": 0.449329137802124, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6142187118530273, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.3531249761581421, "step": 4829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 591.8125, "completions/mean_terminated_length": 563.0, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.966, "grad_norm": 0.3911796510219574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57238970.0, "reward": 0.7567945718765259, "reward_std": 0.4199484586715698, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3885875344276428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7962499856948853, "step": 4830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 739.375, "completions/mean_terminated_length": 720.4000244140625, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.9662, "grad_norm": 0.2054554969072342, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57254736.0, "reward": 0.8583600521087646, "reward_std": 0.403048038482666, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5917187929153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.621874988079071, "step": 4831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 686.9375, "completions/mean_terminated_length": 664.4666748046875, "completions/min_length": 511.0, "completions/min_terminated_length": 511.0, "epoch": 0.9664, "grad_norm": 0.32371434569358826, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57269807.0, "reward": 1.0457944869995117, "reward_std": 0.5087544918060303, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6540875434875488, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.815000057220459, "step": 4832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 838.25, "completions/mean_terminated_length": 776.3333740234375, "completions/min_length": 616.0, "completions/min_terminated_length": 616.0, "epoch": 0.9666, "grad_norm": 0.19876253604888916, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57288243.0, "reward": 0.5548442602157593, "reward_std": 0.37490761280059814, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3596875071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.4625000059604645, "step": 4833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 658.3125, "completions/mean_terminated_length": 573.923095703125, "completions/min_length": 411.0, "completions/min_terminated_length": 411.0, "epoch": 0.9668, "grad_norm": 0.10084617882966995, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57302984.0, "reward": 0.46220386028289795, "reward_std": 0.2520271837711334, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11190624535083771, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7406250238418579, "step": 4834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 701.375, "completions/mean_terminated_length": 679.86669921875, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.967, "grad_norm": 0.23784923553466797, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57318302.0, "reward": 1.018425464630127, "reward_std": 0.38419628143310547, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.8493499755859375, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4112499952316284, "step": 4835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 333.1875, "completions/mean_terminated_length": 333.1875, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.9672, "grad_norm": 0.27006423473358154, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57327057.0, "reward": 0.7086727619171143, "reward_std": 0.32716119289398193, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.29234376549720764, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 520.5, "completions/mean_terminated_length": 520.5, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.9674, "grad_norm": 0.25510215759277344, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57339657.0, "reward": 1.1073977947235107, "reward_std": 0.46912986040115356, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6522937417030334, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9068750143051147, "step": 4837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 803.0625, "completions/mean_terminated_length": 670.5, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.9676, "grad_norm": 0.22398453950881958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57357242.0, "reward": 0.48961004614830017, "reward_std": 0.48511749505996704, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2292187511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.484375, "step": 4838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 475.5, "completions/mean_terminated_length": 475.5, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.9678, "grad_norm": 0.07841335982084274, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57369378.0, "reward": 0.9599221348762512, "reward_std": 0.0774601623415947, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9198437929153442, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.20937499403953552, "step": 4839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 428.375, "completions/mean_terminated_length": 428.375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.968, "grad_norm": 0.18290510773658752, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57381016.0, "reward": 0.8499228954315186, "reward_std": 0.45013052225112915, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38734376430511475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9343749284744263, "step": 4840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 645.4375, "completions/mean_terminated_length": 645.4375, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.9682, "grad_norm": 0.32367590069770813, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57396303.0, "reward": 1.0849883556365967, "reward_std": 0.40519723296165466, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7324750423431396, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7362500429153442, "step": 4841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 514.75, "completions/mean_terminated_length": 480.8000183105469, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.9684, "grad_norm": 0.1082928478717804, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57408715.0, "reward": 1.4053010940551758, "reward_std": 0.37474721670150757, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9355999827384949, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987499475479126, "step": 4842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 452.9375, "completions/mean_terminated_length": 452.9375, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.9686, "grad_norm": 0.1810821294784546, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57420330.0, "reward": 0.9058477878570557, "reward_std": 0.47403061389923096, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4366937577724457, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1008.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 720.375, "completions/mean_terminated_length": 720.375, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.9688, "grad_norm": 0.18760521709918976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57435440.0, "reward": 0.7993756532669067, "reward_std": 0.2799172103404999, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4737499952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6624999642372131, "step": 4844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 593.125, "completions/mean_terminated_length": 564.4000244140625, "completions/min_length": 420.0, "completions/min_terminated_length": 420.0, "epoch": 0.969, "grad_norm": 0.22176815569400787, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57448690.0, "reward": 1.2195758819580078, "reward_std": 0.5120174288749695, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7516499757766724, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987500071525574, "step": 4845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 331.0, "completions/max_terminated_length": 331.0, "completions/mean_length": 287.3125, "completions/mean_terminated_length": 287.3125, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.9692, "grad_norm": 0.4374596178531647, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57457063.0, "reward": 1.3801977634429932, "reward_std": 0.3245638906955719, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8853937387466431, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 309.5625, "completions/mean_terminated_length": 309.5625, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.9694, "grad_norm": 0.4389057457447052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57465904.0, "reward": 1.4389728307724, "reward_std": 0.2373962104320526, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.940443754196167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 372.0625, "completions/mean_terminated_length": 372.0625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.9696, "grad_norm": 0.30456262826919556, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57475825.0, "reward": 1.2594414949417114, "reward_std": 0.4281385540962219, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7688812613487244, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 855.9375, "completions/mean_terminated_length": 779.5454711914062, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.9698, "grad_norm": 0.524406373500824, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57493872.0, "reward": 0.7233474254608154, "reward_std": 0.5672297477722168, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5716937780380249, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.37062501907348633, "step": 4849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 421.5625, "completions/mean_terminated_length": 421.5625, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.97, "grad_norm": 0.3999394476413727, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57504617.0, "reward": 1.088907241821289, "reward_std": 0.4801720380783081, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6153125166893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 497.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 330.625, "completions/mean_terminated_length": 330.625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.9702, "grad_norm": 0.15439949929714203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57513779.0, "reward": 0.9930474758148193, "reward_std": 0.14997844398021698, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.8610937595367432, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.27812498807907104, "step": 4851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 382.625, "completions/mean_terminated_length": 382.625, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.9704, "grad_norm": 0.1901821345090866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57523581.0, "reward": 0.7378132343292236, "reward_std": 0.17449016869068146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4756249785423279, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6749999523162842, "step": 4852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 563.0, "completions/mean_terminated_length": 532.2667236328125, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.9706, "grad_norm": 0.11950165033340454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57536493.0, "reward": 0.5621883273124695, "reward_std": 0.17531739175319672, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18687500059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.831250011920929, "step": 4853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 537.0625, "completions/mean_terminated_length": 504.60003662109375, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.9708, "grad_norm": 0.18944692611694336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57548830.0, "reward": 0.8214067220687866, "reward_std": 0.32289600372314453, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6428124904632568, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.46875, "step": 4854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 599.375, "completions/mean_terminated_length": 599.375, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.971, "grad_norm": 0.24987995624542236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57562276.0, "reward": 1.110859751701355, "reward_std": 0.23006854951381683, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.9717187881469727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.359375, "step": 4855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 456.875, "completions/mean_terminated_length": 456.875, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.9712, "grad_norm": 0.29853832721710205, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57573122.0, "reward": 0.9724228978157043, "reward_std": 0.47904375195503235, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5073437690734863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 452.9375, "completions/mean_terminated_length": 452.9375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.9714, "grad_norm": 0.13994643092155457, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57584657.0, "reward": 0.6417977809906006, "reward_std": 0.2539544701576233, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22109374403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 742.8125, "completions/mean_terminated_length": 649.0833740234375, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.9716, "grad_norm": 0.31043368577957153, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57600846.0, "reward": 0.9355348348617554, "reward_std": 0.6015090346336365, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7460687160491943, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3893750011920929, "step": 4858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 589.625, "completions/mean_terminated_length": 560.6666870117188, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.9718, "grad_norm": 0.25836634635925293, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57614488.0, "reward": 0.7589818239212036, "reward_std": 0.2529570758342743, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5804625153541565, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5399999618530273, "step": 4859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 470.25, "completions/mean_terminated_length": 433.3333435058594, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.972, "grad_norm": 0.5606569647789001, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57625676.0, "reward": 1.0097540616989136, "reward_std": 0.5342838168144226, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5820062160491943, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8518749475479126, "step": 4860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 664.125, "completions/mean_terminated_length": 581.0769653320312, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.9722, "grad_norm": 0.2414819896221161, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57640334.0, "reward": 0.6959379315376282, "reward_std": 0.4132830798625946, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5168750286102295, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4750000238418579, "step": 4861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 457.375, "completions/mean_terminated_length": 457.375, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.9724, "grad_norm": 0.1874600499868393, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57651492.0, "reward": 0.9097539782524109, "reward_std": 0.4015648066997528, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5070062279701233, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7831250429153442, "step": 4862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 454.125, "completions/mean_terminated_length": 454.125, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.9726, "grad_norm": 0.25947389006614685, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57662534.0, "reward": 0.7133604288101196, "reward_std": 0.32949700951576233, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30171874165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.883118748664856, "step": 4863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 256.0, "completions/max_terminated_length": 256.0, "completions/mean_length": 204.0625, "completions/mean_terminated_length": 204.0625, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.9728, "grad_norm": 0.3854849338531494, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57669575.0, "reward": 1.4397321939468384, "reward_std": 0.2375956028699875, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.941962480545044, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500667572021, "step": 4864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 371.375, "completions/mean_terminated_length": 371.375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.973, "grad_norm": 0.20877833664417267, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57679805.0, "reward": 1.0705478191375732, "reward_std": 0.4511719048023224, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6410937309265137, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "step": 4865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 280.0625, "completions/mean_terminated_length": 280.0625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.9732, "grad_norm": 0.4018665850162506, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57688046.0, "reward": 0.777241587638855, "reward_std": 0.4306822717189789, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30448126792907715, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "step": 4866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 626.8125, "completions/mean_terminated_length": 600.3333740234375, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.9734, "grad_norm": 0.1638823300600052, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57701899.0, "reward": 1.2258477210998535, "reward_std": 0.5008319020271301, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7641937732696533, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8831250071525574, "step": 4867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 398.875, "completions/mean_terminated_length": 398.875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.9736, "grad_norm": 0.2976434826850891, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57711961.0, "reward": 0.7396103739738464, "reward_std": 0.3805011510848999, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2917187511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934374988079071, "step": 4868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 970.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 434.75, "completions/mean_terminated_length": 434.75, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.9738, "grad_norm": 0.37727677822113037, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57722565.0, "reward": 0.9624758958816528, "reward_std": 0.4863021969795227, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4874500036239624, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9412499666213989, "step": 4869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 567.4375, "completions/mean_terminated_length": 567.4375, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.974, "grad_norm": 0.16324493288993835, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57735836.0, "reward": 0.6324883699417114, "reward_std": 0.24520243704319, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20247499644756317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "step": 4870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 459.0, "completions/max_terminated_length": 459.0, "completions/mean_length": 356.125, "completions/mean_terminated_length": 356.125, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.9742, "grad_norm": 0.5563121438026428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57746158.0, "reward": 1.3189570903778076, "reward_std": 0.3869653642177582, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8254125118255615, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 656.0, "completions/mean_terminated_length": 631.4666748046875, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.9744, "grad_norm": 0.2632243037223816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57760158.0, "reward": 0.6313289999961853, "reward_std": 0.3597148060798645, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20015624165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8781249523162842, "step": 4872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 667.625, "completions/mean_terminated_length": 643.86669921875, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.9746, "grad_norm": 0.22993150353431702, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57774424.0, "reward": 0.8284383416175842, "reward_std": 0.4489971101284027, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.46937498450279236, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "step": 4873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 905.5, "completions/mean_terminated_length": 813.3333129882812, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.9748, "grad_norm": 0.5076905488967896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57792512.0, "reward": 0.34507861733436584, "reward_std": 0.3222992718219757, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1276562362909317, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.4906249940395355, "step": 4874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 503.6875, "completions/mean_terminated_length": 503.6875, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.975, "grad_norm": 0.25092512369155884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57804347.0, "reward": 0.771094560623169, "reward_std": 0.3202754855155945, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.41718751192092896, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7875000238418579, "step": 4875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 349.0, "completions/mean_terminated_length": 349.0, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.9752, "grad_norm": 0.3024475574493408, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57814011.0, "reward": 1.1964040994644165, "reward_std": 0.4633290469646454, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7053062915802002, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "step": 4876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 402.0625, "completions/mean_terminated_length": 402.0625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.9754, "grad_norm": 0.34798717498779297, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57824492.0, "reward": 0.798488438129425, "reward_std": 0.42116779088974, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3469749987125397, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "step": 4877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 939.8125, "completions/mean_terminated_length": 874.3333129882812, "completions/min_length": 721.0, "completions/min_terminated_length": 721.0, "epoch": 0.9756, "grad_norm": 1.6917587518692017, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57843593.0, "reward": 0.7129692435264587, "reward_std": 0.7018963694572449, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4884375035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.44999998807907104, "step": 4878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 866.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 696.4375, "completions/mean_terminated_length": 696.4375, "completions/min_length": 563.0, "completions/min_terminated_length": 563.0, "epoch": 0.9758, "grad_norm": 0.30653995275497437, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57858672.0, "reward": 0.999457061290741, "reward_std": 0.4303581714630127, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.623912513256073, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7987500429153442, "step": 4879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 335.125, "completions/mean_terminated_length": 335.125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.976, "grad_norm": 0.13930387794971466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57869138.0, "reward": 1.4397697448730469, "reward_std": 0.23760554194450378, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9420374631881714, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "step": 4880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 584.25, "completions/mean_terminated_length": 584.25, "completions/min_length": 439.0, "completions/min_terminated_length": 439.0, "epoch": 0.9762, "grad_norm": 0.05915243923664093, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57882886.0, "reward": 0.5841414928436279, "reward_std": 0.06013471260666847, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16828125715255737, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9093749523162842, "step": 4881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 376.6875, "completions/mean_terminated_length": 376.6875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.9764, "grad_norm": 0.1920018345117569, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57893041.0, "reward": 1.3981883525848389, "reward_std": 0.25957536697387695, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.921375036239624, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8575000166893005, "step": 4882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 278.0, "completions/max_terminated_length": 278.0, "completions/mean_length": 236.3125, "completions/mean_terminated_length": 236.3125, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.9766, "grad_norm": 0.45325371623039246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57900838.0, "reward": 1.3771041631698608, "reward_std": 0.33299002051353455, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8792062997817993, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "step": 4883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 996.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 702.1875, "completions/mean_terminated_length": 702.1875, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.9768, "grad_norm": 0.15056581795215607, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57916409.0, "reward": 1.0464069843292236, "reward_std": 0.35165828466415405, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7803125381469727, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.581250011920929, "step": 4884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 912.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 526.75, "completions/mean_terminated_length": 526.75, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.977, "grad_norm": 0.21821850538253784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57929173.0, "reward": 1.0056252479553223, "reward_std": 0.11906410753726959, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.9487500190734863, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.1875, "step": 4885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 657.1875, "completions/mean_terminated_length": 657.1875, "completions/min_length": 485.0, "completions/min_terminated_length": 485.0, "epoch": 0.9772, "grad_norm": 0.10662835091352463, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57943512.0, "reward": 1.3776321411132812, "reward_std": 0.2690081298351288, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8802624940872192, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8224999904632568, "step": 4886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 159.6875, "completions/mean_terminated_length": 159.6875, "completions/min_length": 98.0, "completions/min_terminated_length": 98.0, "epoch": 0.9774, "grad_norm": 1.952322006225586, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57949715.0, "reward": 1.1911982297897339, "reward_std": 0.4728165566921234, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6948931217193604, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9888125061988831, "step": 4887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 396.625, "completions/mean_terminated_length": 396.625, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.9776, "grad_norm": 0.26764339208602905, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57960173.0, "reward": 0.8413289189338684, "reward_std": 0.3387213945388794, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.49515625834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.784375011920929, "step": 4888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 731.625, "completions/mean_terminated_length": 664.1538696289062, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.9778, "grad_norm": 0.32386669516563416, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57975975.0, "reward": 0.6106257438659668, "reward_std": 0.4346308410167694, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2837499976158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 840.6875, "completions/mean_terminated_length": 757.3636474609375, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.978, "grad_norm": 0.48975637555122375, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57993730.0, "reward": 0.6479690670967102, "reward_std": 0.5526928901672363, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5459374785423279, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.32499998807907104, "step": 4890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 468.0, "completions/max_terminated_length": 468.0, "completions/mean_length": 326.3125, "completions/mean_terminated_length": 326.3125, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.9782, "grad_norm": 0.4202582538127899, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58002775.0, "reward": 1.2216415405273438, "reward_std": 0.41100072860717773, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7557812333106995, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8656250238418579, "step": 4891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 507.4375, "completions/mean_terminated_length": 507.4375, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.9784, "grad_norm": 0.016368618234992027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58014510.0, "reward": 0.5527228116989136, "reward_std": 0.0180070698261261, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10544374585151672, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "step": 4892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 372.0, "completions/mean_terminated_length": 372.0, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.9786, "grad_norm": 0.47061288356781006, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58024494.0, "reward": 1.0953102111816406, "reward_std": 0.43712136149406433, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6906187534332275, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.830625057220459, "step": 4893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 426.75, "completions/mean_terminated_length": 426.75, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.9788, "grad_norm": 0.13365747034549713, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58034938.0, "reward": 1.4373321533203125, "reward_std": 0.24467679858207703, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9371625185012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "step": 4894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 751.0625, "completions/mean_terminated_length": 660.0833740234375, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.979, "grad_norm": 0.16570071876049042, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58051627.0, "reward": 0.6271098852157593, "reward_std": 0.41115039587020874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5042187571525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.31562501192092896, "step": 4895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 379.75, "completions/mean_terminated_length": 379.75, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.9792, "grad_norm": 0.37704557180404663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58061703.0, "reward": 1.1907789707183838, "reward_std": 0.4492145776748657, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6940562725067139, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9181250333786011, "step": 4896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 513.5, "completions/mean_terminated_length": 479.4667053222656, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.9794, "grad_norm": 0.24716298282146454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58073551.0, "reward": 1.0205349922180176, "reward_std": 0.47703108191490173, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6035687327384949, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7143750190734863, "step": 4897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 481.0625, "completions/mean_terminated_length": 481.0625, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.9796, "grad_norm": 0.10224144160747528, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58086016.0, "reward": 1.438204050064087, "reward_std": 0.23723359405994415, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9389062523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "step": 4898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 303.0, "completions/max_terminated_length": 303.0, "completions/mean_length": 232.375, "completions/mean_terminated_length": 232.375, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.9798, "grad_norm": 1.110337734222412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58093206.0, "reward": 1.2868478298187256, "reward_std": 0.39311107993125916, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8236937522888184, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.903124988079071, "step": 4899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 698.0, "completions/max_terminated_length": 698.0, "completions/mean_length": 535.25, "completions/mean_terminated_length": 535.25, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.98, "grad_norm": 0.01801234483718872, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58105882.0, "reward": 0.5680477619171143, "reward_std": 0.04519158601760864, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13609375059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281249642372131, "step": 4900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 800.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 481.5625, "completions/mean_terminated_length": 481.5625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.9802, "grad_norm": 0.14395886659622192, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58117251.0, "reward": 0.8552349209785461, "reward_std": 0.15304407477378845, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7104687690734863, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4281249940395355, "step": 4901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 640.375, "completions/mean_terminated_length": 614.800048828125, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.9804, "grad_norm": 0.07661435753107071, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58131945.0, "reward": 0.8271098136901855, "reward_std": 0.2690281569957733, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7167187333106995, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.34062501788139343, "step": 4902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 389.3125, "completions/mean_terminated_length": 389.3125, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.9806, "grad_norm": 0.2486814260482788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58142606.0, "reward": 0.9442914128303528, "reward_std": 0.34349116683006287, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7010812759399414, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5456249713897705, "step": 4903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 674.0, "completions/max_terminated_length": 674.0, "completions/mean_length": 446.3125, "completions/mean_terminated_length": 446.3125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.9808, "grad_norm": 0.33138227462768555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58153939.0, "reward": 1.2840633392333984, "reward_std": 0.38930749893188477, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8181250095367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9125000238418579, "step": 4904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 518.0625, "completions/mean_terminated_length": 518.0625, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.981, "grad_norm": 0.11154177784919739, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58165924.0, "reward": 1.4391415119171143, "reward_std": 0.2374393343925476, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9407812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 544.4375, "completions/mean_terminated_length": 512.4666748046875, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.9812, "grad_norm": 0.27672141790390015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58178955.0, "reward": 0.6778913736343384, "reward_std": 0.2679133713245392, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4182812571525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6031250357627869, "step": 4906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 258.5625, "completions/mean_terminated_length": 258.5625, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 0.9814, "grad_norm": 0.7049764394760132, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58187060.0, "reward": 0.7718884944915771, "reward_std": 0.4340342581272125, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2937749922275543, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 4907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 643.8125, "completions/mean_terminated_length": 618.4666748046875, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.9816, "grad_norm": 0.11303810030221939, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58201537.0, "reward": 0.8185162544250488, "reward_std": 0.36034950613975525, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.5745312571525574, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5718749761581421, "step": 4908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 505.75, "completions/mean_terminated_length": 505.75, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.9818, "grad_norm": 0.03575127571821213, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58214317.0, "reward": 0.9712502360343933, "reward_std": 0.03989557549357414, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9424999952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.20000000298023224, "step": 4909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 620.4375, "completions/mean_terminated_length": 620.4375, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.982, "grad_norm": 0.10285256803035736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58228052.0, "reward": 0.8000786900520325, "reward_std": 0.1267356425523758, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6001562476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6031250357627869, "step": 4910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 417.5, "completions/mean_terminated_length": 417.5, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.9822, "grad_norm": 0.17106474936008453, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58239212.0, "reward": 0.6596883535385132, "reward_std": 0.2557787299156189, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.25687500834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 4911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 461.125, "completions/mean_terminated_length": 461.125, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.9824, "grad_norm": 0.00022402184549719095, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58250238.0, "reward": 1.4989571571350098, "reward_std": 0.00044457349576987326, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9979125261306763, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "step": 4912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 677.0, "completions/mean_terminated_length": 596.923095703125, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.9826, "grad_norm": 0.2227955460548401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58264990.0, "reward": 0.8821097612380981, "reward_std": 0.49397754669189453, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.4517187476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.34062498807907104, "step": 4913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 681.5625, "completions/mean_terminated_length": 681.5625, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.9828, "grad_norm": 0.10294032096862793, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58280823.0, "reward": 0.7655414938926697, "reward_std": 0.24068857729434967, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5310811996459961, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5018749833106995, "step": 4914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 247.75, "completions/mean_terminated_length": 247.75, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.983, "grad_norm": 0.0006850389181636274, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58288371.0, "reward": 1.4994947910308838, "reward_std": 0.0005964109441265464, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9989875555038452, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "step": 4915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 584.125, "completions/mean_terminated_length": 584.125, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.9832, "grad_norm": 0.17383894324302673, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58302021.0, "reward": 0.7794539332389832, "reward_std": 0.2923417389392853, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4339062571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7281249761581421, "step": 4916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 352.5625, "completions/mean_terminated_length": 352.5625, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.9834, "grad_norm": 0.2511637806892395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58311726.0, "reward": 1.4394415616989136, "reward_std": 0.23751763999462128, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9413812160491943, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 322.5625, "completions/mean_terminated_length": 322.5625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.9836, "grad_norm": 0.6079227328300476, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58320311.0, "reward": 1.257922887802124, "reward_std": 0.43156465888023376, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7658437490463257, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "step": 4918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 547.875, "completions/mean_terminated_length": 547.875, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.9838, "grad_norm": 0.1483164131641388, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58333765.0, "reward": 0.8442972898483276, "reward_std": 0.195061594247818, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6885937452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4156250059604645, "step": 4919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 357.0, "completions/mean_terminated_length": 357.0, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.984, "grad_norm": 0.012113607488572598, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58343189.0, "reward": 0.5445165634155273, "reward_std": 0.011709420010447502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "step": 4920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 471.0, "completions/max_terminated_length": 471.0, "completions/mean_length": 341.6875, "completions/mean_terminated_length": 341.6875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.9842, "grad_norm": 0.3703450560569763, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58352160.0, "reward": 1.3212261199951172, "reward_std": 0.38325533270835876, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8299499750137329, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "step": 4921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 383.8125, "completions/mean_terminated_length": 383.8125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.9844, "grad_norm": 0.35171863436698914, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58362605.0, "reward": 0.7995321750640869, "reward_std": 0.4185425639152527, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3490625023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 4922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 444.375, "completions/mean_terminated_length": 444.375, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.9846, "grad_norm": 0.5162942409515381, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58373731.0, "reward": 0.7920321822166443, "reward_std": 0.4216875433921814, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33406251668930054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "step": 4923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 467.5, "completions/mean_terminated_length": 467.5, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.9848, "grad_norm": 0.11207836866378784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58385243.0, "reward": 0.884297251701355, "reward_std": 0.17106129229068756, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7685937881469727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.34687498211860657, "step": 4924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 448.625, "completions/mean_terminated_length": 448.625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.985, "grad_norm": 0.4252547323703766, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58396885.0, "reward": 0.8823444843292236, "reward_std": 0.3623858392238617, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5771875381469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.668749988079071, "step": 4925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 625.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 415.25, "completions/mean_terminated_length": 415.25, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.9852, "grad_norm": 0.2775042951107025, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58407833.0, "reward": 1.26414155960083, "reward_std": 0.41980940103530884, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7782812118530273, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "step": 4926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 488.1875, "completions/mean_terminated_length": 452.4667053222656, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.9854, "grad_norm": 0.5119762420654297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58419868.0, "reward": 0.7026569843292236, "reward_std": 0.37224575877189636, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.34281250834465027, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "step": 4927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 527.5, "completions/mean_terminated_length": 494.4000244140625, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.9856, "grad_norm": 0.23505793511867523, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58432884.0, "reward": 0.5971821546554565, "reward_std": 0.3025156259536743, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19436250627040863, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8362500071525574, "step": 4928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 630.25, "completions/mean_terminated_length": 574.0, "completions/min_length": 435.0, "completions/min_terminated_length": 435.0, "epoch": 0.9858, "grad_norm": 0.29216256737709045, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58447208.0, "reward": 1.1607788801193237, "reward_std": 0.5611714720726013, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7590562105178833, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7743750214576721, "step": 4929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 375.9375, "completions/mean_terminated_length": 375.9375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.986, "grad_norm": 0.21585534512996674, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58457095.0, "reward": 1.4085698127746582, "reward_std": 0.2609133720397949, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.9421374797821045, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9037500023841858, "step": 4930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 615.9375, "completions/mean_terminated_length": 615.9375, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.9862, "grad_norm": 0.16866837441921234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58471302.0, "reward": 1.1496883630752563, "reward_std": 0.4257732331752777, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.736875057220459, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8499999642372131, "step": 4931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 744.25, "completions/mean_terminated_length": 679.6923217773438, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.9864, "grad_norm": 0.41569551825523376, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58489082.0, "reward": 0.7764720916748047, "reward_std": 0.40272051095962524, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.740443766117096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.14562499523162842, "step": 4932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 438.0, "completions/max_terminated_length": 438.0, "completions/mean_length": 322.3125, "completions/mean_terminated_length": 322.3125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.9866, "grad_norm": 0.381817489862442, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58497919.0, "reward": 1.4398634433746338, "reward_std": 0.2376306653022766, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9422250390052795, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "step": 4933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 520.0, "completions/mean_terminated_length": 486.4000244140625, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.9868, "grad_norm": 0.1585514396429062, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58510543.0, "reward": 1.1922165155410767, "reward_std": 0.5024663805961609, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7594313025474548, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8543750047683716, "step": 4934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 288.5, "completions/mean_terminated_length": 288.5, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.987, "grad_norm": 0.39759543538093567, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58519623.0, "reward": 1.081382155418396, "reward_std": 0.4889579117298126, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6002625226974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 527.25, "completions/mean_terminated_length": 494.13336181640625, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.9872, "grad_norm": 0.4247090816497803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58532123.0, "reward": 0.8710945248603821, "reward_std": 0.4472842514514923, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5546875, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6812500357627869, "step": 4936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 750.5, "completions/mean_terminated_length": 732.2667236328125, "completions/min_length": 489.0, "completions/min_terminated_length": 489.0, "epoch": 0.9874, "grad_norm": 0.1327957659959793, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58548355.0, "reward": 0.760678768157959, "reward_std": 0.3656542897224426, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4588562250137329, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6506249904632568, "step": 4937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 233.625, "completions/mean_terminated_length": 233.625, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.9876, "grad_norm": 0.011015729047358036, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58556749.0, "reward": 0.9975006580352783, "reward_std": 0.005164253991097212, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.9950000047683716, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.03750000149011612, "step": 4938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 529.25, "completions/mean_terminated_length": 529.25, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.9878, "grad_norm": 0.021423792466521263, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58571985.0, "reward": 0.974219024181366, "reward_std": 0.03997354581952095, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.948437511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.18124999105930328, "step": 4939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 434.625, "completions/mean_terminated_length": 434.625, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.988, "grad_norm": 0.44675135612487793, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58582571.0, "reward": 1.2145321369171143, "reward_std": 0.4374695122241974, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7415624856948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "step": 4940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 442.8125, "completions/mean_terminated_length": 442.8125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.9882, "grad_norm": 0.19209016859531403, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58593256.0, "reward": 0.7560164332389832, "reward_std": 0.31912708282470703, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3870312571525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8218749761581421, "step": 4941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 278.3125, "completions/mean_terminated_length": 278.3125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.9884, "grad_norm": 0.00032074967748485506, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58601693.0, "reward": 1.499035358428955, "reward_std": 0.0005109267658554018, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980687499046326, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "step": 4942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 818.1875, "completions/mean_terminated_length": 770.6923217773438, "completions/min_length": 538.0, "completions/min_terminated_length": 538.0, "epoch": 0.9886, "grad_norm": 0.25372034311294556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58618768.0, "reward": 0.7042972445487976, "reward_std": 0.3588414490222931, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5960937738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.3843750059604645, "step": 4943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 323.0, "completions/max_terminated_length": 323.0, "completions/mean_length": 206.875, "completions/mean_terminated_length": 206.875, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.9888, "grad_norm": 0.9213141798973083, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58625774.0, "reward": 1.2645072937011719, "reward_std": 0.4200277030467987, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7790125012397766, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 702.875, "completions/mean_terminated_length": 657.0, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.989, "grad_norm": 0.12226295471191406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58641932.0, "reward": 0.5313289165496826, "reward_std": 0.24564403295516968, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1876562535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.746874988079071, "step": 4945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 350.0, "completions/max_terminated_length": 350.0, "completions/mean_length": 279.375, "completions/mean_terminated_length": 279.375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.9892, "grad_norm": 0.24796579778194427, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58650802.0, "reward": 0.6657571792602539, "reward_std": 0.32545217871665955, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20651249587535858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 432.0, "completions/max_terminated_length": 432.0, "completions/mean_length": 378.75, "completions/mean_terminated_length": 378.75, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.9894, "grad_norm": 0.14456409215927124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58660670.0, "reward": 1.4397227764129639, "reward_std": 0.23759308457374573, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9419437646865845, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "step": 4947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 571.0, "completions/max_terminated_length": 571.0, "completions/mean_length": 412.6875, "completions/mean_terminated_length": 412.6875, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.9896, "grad_norm": 0.2206587791442871, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58671881.0, "reward": 0.850156843662262, "reward_std": 0.21821610629558563, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7003124952316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.375, "step": 4948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 677.375, "completions/mean_terminated_length": 627.857177734375, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.9898, "grad_norm": 0.2051655501127243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58687695.0, "reward": 0.6758600473403931, "reward_std": 0.40632426738739014, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3517187535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.703125, "step": 4949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 709.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 531.9375, "completions/mean_terminated_length": 531.9375, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.99, "grad_norm": 0.16245263814926147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58699838.0, "reward": 1.099219560623169, "reward_std": 0.38246628642082214, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.760937511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7250000238418579, "step": 4950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 201.0, "completions/max_terminated_length": 201.0, "completions/mean_length": 166.8125, "completions/mean_terminated_length": 166.8125, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "epoch": 0.9902, "grad_norm": 0.984749436378479, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58706107.0, "reward": 1.0115292072296143, "reward_std": 0.5044964551925659, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5230562686920166, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "step": 4951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 476.0, "completions/mean_terminated_length": 476.0, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.9904, "grad_norm": 0.36056819558143616, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58717307.0, "reward": 0.9674103260040283, "reward_std": 0.48439666628837585, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4973187744617462, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "step": 4952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 383.375, "completions/mean_terminated_length": 383.375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.9906, "grad_norm": 0.11669478565454483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58726977.0, "reward": 0.6383476257324219, "reward_std": 0.1554064005613327, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.276693731546402, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8143750429153442, "step": 4953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 440.3125, "completions/mean_terminated_length": 440.3125, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.9908, "grad_norm": 0.11714527755975723, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58738374.0, "reward": 0.6608603000640869, "reward_std": 0.24761216342449188, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2592187523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.878125011920929, "step": 4954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 360.0, "completions/max_terminated_length": 360.0, "completions/mean_length": 226.5625, "completions/mean_terminated_length": 226.5625, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.991, "grad_norm": 0.6267589926719666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58746367.0, "reward": 1.322550892829895, "reward_std": 0.3552817106246948, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8325999975204468, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "step": 4955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 290.0, "completions/max_terminated_length": 290.0, "completions/mean_length": 245.0625, "completions/mean_terminated_length": 245.0625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.9912, "grad_norm": 0.14406053721904755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58754432.0, "reward": 0.6272633075714111, "reward_std": 0.1285761445760727, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25452500581741333, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8462499976158142, "step": 4956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 955.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 712.0, "completions/mean_terminated_length": 712.0, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.9914, "grad_norm": 0.07639224827289581, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58769712.0, "reward": 1.092344045639038, "reward_std": 0.2211698591709137, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.9346874952316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3500000238418579, "step": 4957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 429.4375, "completions/mean_terminated_length": 389.8000183105469, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.9916, "grad_norm": 0.0704599991440773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58780663.0, "reward": 0.5373983979225159, "reward_std": 0.28642505407333374, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0747937485575676, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9306249618530273, "step": 4958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 733.6875, "completions/mean_terminated_length": 714.3333740234375, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.9918, "grad_norm": 0.5135071873664856, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58797074.0, "reward": 0.9988288283348083, "reward_std": 0.45642754435539246, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6851562261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6781250238418579, "step": 4959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 672.625, "completions/mean_terminated_length": 649.2000122070312, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.992, "grad_norm": 0.364347368478775, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58811980.0, "reward": 1.2660164833068848, "reward_std": 0.46806275844573975, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.782031238079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "step": 4960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 435.9375, "completions/mean_terminated_length": 435.9375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.9922, "grad_norm": 0.2344200164079666, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58822651.0, "reward": 0.8426447510719299, "reward_std": 0.45715752243995667, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3727875053882599, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "step": 4961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 369.25, "completions/mean_terminated_length": 369.25, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.9924, "grad_norm": 0.5314513444900513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58832799.0, "reward": 1.1803133487701416, "reward_std": 0.43270397186279297, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7356250286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "step": 4962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 451.3125, "completions/mean_terminated_length": 451.3125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.9926, "grad_norm": 0.22915273904800415, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58844356.0, "reward": 1.0100635290145874, "reward_std": 0.490188330411911, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5201249718666077, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9325000047683716, "step": 4963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 403.3125, "completions/mean_terminated_length": 403.3125, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.9928, "grad_norm": 0.19858123362064362, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58855033.0, "reward": 0.7168758511543274, "reward_std": 0.32588720321655273, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3087499737739563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8687499761581421, "step": 4964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 345.1875, "completions/mean_terminated_length": 345.1875, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.993, "grad_norm": 0.00035785743966698647, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58864140.0, "reward": 1.4990322589874268, "reward_std": 0.0005046323640272021, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9980624914169312, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "step": 4965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 537.8125, "completions/mean_terminated_length": 537.8125, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.9932, "grad_norm": 0.21050868928432465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58877161.0, "reward": 0.9762508869171143, "reward_std": 0.4035230278968811, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6399999856948853, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 351.125, "completions/mean_terminated_length": 306.2666931152344, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.9934, "grad_norm": 0.7299486398696899, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58887307.0, "reward": 1.0705602169036865, "reward_std": 0.5933318734169006, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6411187648773193, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9025000333786011, "step": 4967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 499.0, "completions/max_terminated_length": 499.0, "completions/mean_length": 274.25, "completions/mean_terminated_length": 274.25, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.9936, "grad_norm": 0.21803049743175507, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58895679.0, "reward": 0.6606072187423706, "reward_std": 0.32675665616989136, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.196212500333786, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "step": 4968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 821.75, "completions/mean_terminated_length": 729.8181762695312, "completions/min_length": 483.0, "completions/min_terminated_length": 483.0, "epoch": 0.9938, "grad_norm": 0.37420955300331116, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58913067.0, "reward": 0.6238161325454712, "reward_std": 0.5097531080245972, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4351312518119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.47062498331069946, "step": 4969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 680.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 348.25, "completions/mean_terminated_length": 348.25, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.994, "grad_norm": 0.41052302718162537, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58922415.0, "reward": 1.2896759510040283, "reward_std": 0.3845372796058655, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8293499946594238, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9112499952316284, "step": 4970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 764.6875, "completions/mean_terminated_length": 747.4000244140625, "completions/min_length": 457.0, "completions/min_terminated_length": 457.0, "epoch": 0.9942, "grad_norm": 0.17362692952156067, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58938586.0, "reward": 0.7145318388938904, "reward_std": 0.28395184874534607, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.49156248569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5249999761581421, "step": 4971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 572.0, "completions/max_terminated_length": 572.0, "completions/mean_length": 389.8125, "completions/mean_terminated_length": 389.8125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.9944, "grad_norm": 0.3203250765800476, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58948999.0, "reward": 1.2692853212356567, "reward_std": 0.4111281633377075, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7885687947273254, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9393749833106995, "step": 4972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 716.0, "completions/mean_terminated_length": 644.923095703125, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.9946, "grad_norm": 0.07743438333272934, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58965207.0, "reward": 0.4937506914138794, "reward_std": 0.2599954903125763, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17499999701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7124999761581421, "step": 4973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 980.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 546.0, "completions/mean_terminated_length": 546.0, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.9948, "grad_norm": 0.2671853005886078, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58978695.0, "reward": 0.8821102380752563, "reward_std": 0.44112348556518555, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4517187476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "step": 4974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 980.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 721.4375, "completions/mean_terminated_length": 721.4375, "completions/min_length": 583.0, "completions/min_terminated_length": 583.0, "epoch": 0.995, "grad_norm": 0.09099411219358444, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58994286.0, "reward": 0.7216413021087646, "reward_std": 0.18351495265960693, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.44328123331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.671875, "step": 4975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 383.875, "completions/mean_terminated_length": 383.875, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.9952, "grad_norm": 0.37725621461868286, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59004300.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 514.25, "completions/mean_terminated_length": 514.25, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.9954, "grad_norm": 0.13853102922439575, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59016832.0, "reward": 0.7692194581031799, "reward_std": 0.27253803610801697, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.47593748569488525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6749999523162842, "step": 4977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 773.25, "completions/mean_terminated_length": 756.5333862304688, "completions/min_length": 525.0, "completions/min_terminated_length": 525.0, "epoch": 0.9956, "grad_norm": 0.2250785380601883, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59033524.0, "reward": 0.7156260013580322, "reward_std": 0.4169578552246094, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3062500059604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "step": 4978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 307.0, "completions/max_terminated_length": 307.0, "completions/mean_length": 249.0625, "completions/mean_terminated_length": 249.0625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.9958, "grad_norm": 0.2495373785495758, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59041557.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 995.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 584.1875, "completions/mean_terminated_length": 584.1875, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.996, "grad_norm": 0.16128946840763092, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59055128.0, "reward": 0.6271855235099792, "reward_std": 0.1632511168718338, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25436875224113464, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.830625057220459, "step": 4980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 357.875, "completions/mean_terminated_length": 357.875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.9962, "grad_norm": 0.7532382011413574, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59065622.0, "reward": 1.1718757152557373, "reward_std": 0.4152379631996155, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.71875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.793749988079071, "step": 4981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 440.4375, "completions/mean_terminated_length": 440.4375, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.9964, "grad_norm": 0.4437426030635834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59077437.0, "reward": 1.3773759603500366, "reward_std": 0.33309629559516907, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8797500133514404, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "step": 4982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 413.0, "completions/max_terminated_length": 413.0, "completions/mean_length": 330.3125, "completions/mean_terminated_length": 330.3125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.9966, "grad_norm": 0.18958988785743713, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59086706.0, "reward": 1.439375877380371, "reward_std": 0.23749998211860657, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9412499666213989, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 430.0, "completions/max_terminated_length": 430.0, "completions/mean_length": 283.125, "completions/mean_terminated_length": 283.125, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.9968, "grad_norm": 0.3170463740825653, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59095012.0, "reward": 1.351525902748108, "reward_std": 0.32066813111305237, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8905500173568726, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8774999380111694, "step": 4984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 414.9375, "completions/mean_terminated_length": 414.9375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.997, "grad_norm": 0.24833187460899353, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59106259.0, "reward": 0.8024759292602539, "reward_std": 0.4209013879299164, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3549499809741974, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9287499785423279, "step": 4985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 237.875, "completions/mean_terminated_length": 237.875, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 0.9972, "grad_norm": 0.5566157698631287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59113441.0, "reward": 0.9590446949005127, "reward_std": 0.49208420515060425, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48058751225471497, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "step": 4986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 488.0625, "completions/mean_terminated_length": 488.0625, "completions/min_length": 397.0, "completions/min_terminated_length": 397.0, "epoch": 0.9974, "grad_norm": 0.11210137605667114, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59125154.0, "reward": 1.4397790431976318, "reward_std": 0.23760807514190674, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9420562386512756, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "step": 4987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 655.3125, "completions/mean_terminated_length": 630.7333374023438, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.9976, "grad_norm": 0.16047802567481995, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59139367.0, "reward": 0.6316413879394531, "reward_std": 0.243261456489563, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3257812559604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7093750238418579, "step": 4988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 479.6875, "completions/mean_terminated_length": 479.6875, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.9978, "grad_norm": 0.14864283800125122, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59150850.0, "reward": 0.7306256890296936, "reward_std": 0.1633245348930359, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4612500071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6812499761581421, "step": 4989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 848.875, "completions/mean_terminated_length": 823.857177734375, "completions/min_length": 517.0, "completions/min_terminated_length": 517.0, "epoch": 0.998, "grad_norm": 0.40608295798301697, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59168656.0, "reward": 1.0087473392486572, "reward_std": 0.46849486231803894, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.8299937844276428, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4181249737739563, "step": 4990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 842.0, "completions/max_terminated_length": 842.0, "completions/mean_length": 588.875, "completions/mean_terminated_length": 588.875, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.9982, "grad_norm": 0.20117440819740295, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59182062.0, "reward": 0.9313946962356567, "reward_std": 0.467390239238739, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.48778748512268066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8924999833106995, "step": 4991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 616.0, "completions/max_terminated_length": 616.0, "completions/mean_length": 417.375, "completions/mean_terminated_length": 417.375, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.9984, "grad_norm": 0.33665600419044495, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59192580.0, "reward": 0.8712508678436279, "reward_std": 0.43613019585609436, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.429999977350235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "step": 4992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 702.8125, "completions/mean_terminated_length": 656.9285888671875, "completions/min_length": 505.0, "completions/min_terminated_length": 505.0, "epoch": 0.9986, "grad_norm": 0.16504128277301788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59209265.0, "reward": 0.7089848518371582, "reward_std": 0.37183254957199097, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.48046875, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.546875, "step": 4993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 670.9375, "completions/mean_terminated_length": 670.9375, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.9988, "grad_norm": 0.1403464376926422, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59224400.0, "reward": 0.7633476257324219, "reward_std": 0.27947238087654114, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.464193731546402, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6831250190734863, "step": 4994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 908.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 636.1875, "completions/mean_terminated_length": 636.1875, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.999, "grad_norm": 0.10211227834224701, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59238355.0, "reward": 0.7688288688659668, "reward_std": 0.27133458852767944, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4751562476158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.690625011920929, "step": 4995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 354.375, "completions/mean_terminated_length": 354.375, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.9992, "grad_norm": 0.08276461809873581, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59248089.0, "reward": 1.439450979232788, "reward_std": 0.23752018809318542, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9414000511169434, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "step": 4996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 351.6875, "completions/mean_terminated_length": 351.6875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.9994, "grad_norm": 0.61421799659729, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59257924.0, "reward": 1.023750901222229, "reward_std": 0.4905778765678406, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "step": 4997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 658.25, "completions/mean_terminated_length": 658.25, "completions/min_length": 511.0, "completions/min_terminated_length": 511.0, "epoch": 0.9996, "grad_norm": 0.13124850392341614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 59272632.0, "reward": 0.9313287734985352, "reward_std": 0.33127036690711975, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.6751562356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5968749523162842, "step": 4998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 420.625, "completions/mean_terminated_length": 420.625, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.9998, "grad_norm": 0.44137921929359436, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59283522.0, "reward": 0.9802321195602417, "reward_std": 0.4750889241695404, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5229625105857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9337500333786011, "step": 4999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 422.9375, "completions/mean_terminated_length": 382.86669921875, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 1.0, "grad_norm": 0.5364050269126892, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 59294241.0, "reward": 0.8746570944786072, "reward_std": 0.5177679061889648, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43681249022483826, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "step": 5000 }, { "epoch": 1.0, "step": 5000, "total_flos": 0.0, "train_loss": 7.89857585914433e-10, "train_runtime": 100713.2183, "train_samples_per_second": 0.05, "train_steps_per_second": 0.05 } ], "logging_steps": 1, "max_steps": 5000, "num_input_tokens_seen": 59294241, "num_train_epochs": 1, "save_steps": 250, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }