{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 1.7504, "eval_steps": 500, "global_step": 2188, "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": 846.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 509.375, "completions/mean_terminated_length": 509.375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.0002, "grad_norm": 0.07728996127843857, "learning_rate": 0.0, "loss": -0.0, "num_tokens": 12438.0, "reward": 0.5994540452957153, "reward_std": 0.6872168779373169, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3864062428474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7718750238418579, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 966.0, "completions/mean_length": 723.0, "completions/mean_terminated_length": 680.0, "completions/min_length": 417.0, "completions/min_terminated_length": 417.0, "epoch": 0.0004, "grad_norm": 0.04167918115854263, "learning_rate": 4e-09, "loss": 0.0, "num_tokens": 28678.0, "reward": 0.1556258499622345, "reward_std": 0.3504798412322998, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1237500011920929, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.46875, "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": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 453.0, "completions/mean_terminated_length": 453.0, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.0006, "grad_norm": 0.04909047856926918, "learning_rate": 8e-09, "loss": -0.0, "num_tokens": 40422.0, "reward": 0.16234171390533447, "reward_std": 0.40252792835235596, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07468125224113464, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9668749570846558, "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": 268.0, "completions/max_terminated_length": 268.0, "completions/mean_length": 194.3125, "completions/mean_terminated_length": 194.3125, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.0008, "grad_norm": 0.1323743313550949, "learning_rate": 1.2e-08, "loss": 0.0, "num_tokens": 47499.0, "reward": 1.1245856285095215, "reward_std": 0.6705726385116577, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7491687536239624, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9787500500679016, "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": 814.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 512.6875, "completions/mean_terminated_length": 512.6875, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.001, "grad_norm": 0.03757597878575325, "learning_rate": 1.6e-08, "loss": -0.0, "num_tokens": 59558.0, "reward": 0.19240760803222656, "reward_std": 0.25678080320358276, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.009812499396502972, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8574999570846558, "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": 0.0, "completions/max_length": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 524.0, "completions/mean_terminated_length": 524.0, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0012, "grad_norm": 0.049602970480918884, "learning_rate": 2e-08, "loss": -0.0, "num_tokens": 72422.0, "reward": 0.1905481219291687, "reward_std": 0.4449867010116577, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13109374046325684, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.8243750333786011, "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": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 376.6875, "completions/mean_terminated_length": 376.6875, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.0014, "grad_norm": 0.1060582846403122, "learning_rate": 2.4e-08, "loss": 0.0, "num_tokens": 81937.0, "reward": 0.7619545459747314, "reward_std": 0.6273807883262634, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39890623092651367, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9768750071525574, "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": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 398.5, "completions/mean_terminated_length": 398.5, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.0016, "grad_norm": 0.09238973259925842, "learning_rate": 2.8e-08, "loss": -0.0, "num_tokens": 92233.0, "reward": 0.7478917837142944, "reward_std": 0.6964043974876404, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4332812428474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.934374988079071, "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": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 544.0, "completions/mean_terminated_length": 544.0, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.0018, "grad_norm": 0.06984735280275345, "learning_rate": 3.2e-08, "loss": 0.0, "num_tokens": 104841.0, "reward": 0.3294887840747833, "reward_std": 0.6111865639686584, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2214750051498413, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8668749928474426, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 628.9375, "completions/mean_terminated_length": 602.6000366210938, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.002, "grad_norm": 0.020992958918213844, "learning_rate": 3.6e-08, "loss": 0.0, "num_tokens": 119080.0, "reward": 0.05468861013650894, "reward_std": 0.21874983608722687, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.046875, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.7993749976158142, "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": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 313.0625, "completions/mean_terminated_length": 313.0625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.0022, "grad_norm": 0.11147095263004303, "learning_rate": 4e-08, "loss": -0.0, "num_tokens": 128057.0, "reward": 0.6039479374885559, "reward_std": 0.6622610688209534, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3328937292098999, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9018750190734863, "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": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 339.75, "completions/mean_terminated_length": 339.75, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.0024, "grad_norm": 0.0986943319439888, "learning_rate": 4.4e-08, "loss": 0.0, "num_tokens": 137173.0, "reward": 0.5965479016304016, "reward_std": 0.7346310615539551, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38059377670288086, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 0.0, "completions/max_length": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 295.4375, "completions/mean_terminated_length": 295.4375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.0026, "grad_norm": 0.05390046536922455, "learning_rate": 4.8e-08, "loss": -0.0, "num_tokens": 145980.0, "reward": 0.1623421013355255, "reward_std": 0.40252774953842163, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07468124479055405, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9887499809265137, "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": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 681.0625, "completions/mean_terminated_length": 681.0625, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.0028, "grad_norm": 0.02850896306335926, "learning_rate": 5.1999999999999996e-08, "loss": 0.0, "num_tokens": 160605.0, "reward": 0.11984504759311676, "reward_std": 0.26392561197280884, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.052187494933605194, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.7962499856948853, "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": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 366.25, "completions/mean_terminated_length": 366.25, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.003, "grad_norm": 0.05502406135201454, "learning_rate": 5.6e-08, "loss": 0.0, "num_tokens": 170897.0, "reward": 0.25911378860473633, "reward_std": 0.41430312395095825, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08072499930858612, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9187500476837158, "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": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 352.1875, "completions/mean_terminated_length": 352.1875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.0032, "grad_norm": 0.09584761410951614, "learning_rate": 6e-08, "loss": 0.0, "num_tokens": 181220.0, "reward": 0.9384857416152954, "reward_std": 0.6790188550949097, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5644687414169312, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9768750071525574, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 739.4375, "completions/mean_terminated_length": 698.7857666015625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.0034, "grad_norm": 0.03605782240629196, "learning_rate": 6.4e-08, "loss": -0.0, "num_tokens": 198011.0, "reward": 0.21647268533706665, "reward_std": 0.3472157120704651, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12044374644756317, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.5112500190734863, "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": 0.0, "completions/max_length": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 368.4375, "completions/mean_terminated_length": 368.4375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.0036, "grad_norm": 0.09954540431499481, "learning_rate": 6.8e-08, "loss": -0.0, "num_tokens": 207970.0, "reward": 0.46851038932800293, "reward_std": 0.71770179271698, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3120187520980835, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 404.125, "completions/mean_terminated_length": 404.125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.0038, "grad_norm": 0.03911694884300232, "learning_rate": 7.2e-08, "loss": 0.0, "num_tokens": 218452.0, "reward": 0.3723040819168091, "reward_std": 0.25995582342147827, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.057106249034404755, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 0.0, "completions/max_length": 764.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 560.0, "completions/mean_terminated_length": 560.0, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.004, "grad_norm": 0.04315074533224106, "learning_rate": 7.599999999999999e-08, "loss": -0.0, "num_tokens": 233332.0, "reward": 0.1567978858947754, "reward_std": 0.3491474390029907, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12609374523162842, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.640625, "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": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 344.75, "completions/mean_terminated_length": 344.75, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0042, "grad_norm": 0.10559267550706863, "learning_rate": 8e-08, "loss": 0.0, "num_tokens": 242512.0, "reward": 0.7618762850761414, "reward_std": 0.6272827982902527, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.398749977350235, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.90625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 426.625, "completions/mean_terminated_length": 386.8000183105469, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.0044, "grad_norm": 0.0388636589050293, "learning_rate": 8.4e-08, "loss": -0.0, "num_tokens": 253738.0, "reward": 0.16843858361244202, "reward_std": 0.2582721412181854, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02437499910593033, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.840624988079071, "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": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 605.0625, "completions/mean_terminated_length": 605.0625, "completions/min_length": 381.0, "completions/min_terminated_length": 381.0, "epoch": 0.0046, "grad_norm": 0.0358256921172142, "learning_rate": 8.8e-08, "loss": -0.0, "num_tokens": 267195.0, "reward": 0.4082820415496826, "reward_std": 0.33879315853118896, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1915625035762787, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.734375, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 519.0625, "completions/mean_terminated_length": 485.4000244140625, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.0048, "grad_norm": 0.04170688986778259, "learning_rate": 9.199999999999999e-08, "loss": 0.0, "num_tokens": 280060.0, "reward": 0.2503729462623596, "reward_std": 0.30145105719566345, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06324374675750732, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8849999904632568, "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": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 292.4375, "completions/mean_terminated_length": 292.4375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.005, "grad_norm": 0.09910587221384048, "learning_rate": 9.6e-08, "loss": -0.0, "num_tokens": 288483.0, "reward": 0.9443604350090027, "reward_std": 0.6748220920562744, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5762187242507935, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9068750143051147, "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": 910.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 534.4375, "completions/mean_terminated_length": 534.4375, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.0052, "grad_norm": 0.04074499011039734, "learning_rate": 1e-07, "loss": 0.0, "num_tokens": 300714.0, "reward": 0.22538898885250092, "reward_std": 0.41196751594543457, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07577499747276306, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9818750023841858, "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": 438.0, "completions/max_terminated_length": 438.0, "completions/mean_length": 327.375, "completions/mean_terminated_length": 327.375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.0054, "grad_norm": 0.06818709522485733, "learning_rate": 1.0399999999999999e-07, "loss": -0.0, "num_tokens": 309664.0, "reward": 0.31242358684539795, "reward_std": 0.5121544003486633, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12484374642372131, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.987500011920929, "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": 0.0, "completions/max_length": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 395.1875, "completions/mean_terminated_length": 395.1875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.0056, "grad_norm": 0.08183956146240234, "learning_rate": 1.0799999999999999e-07, "loss": -0.0, "num_tokens": 320083.0, "reward": 0.5232821702957153, "reward_std": 0.48883894085884094, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4840624928474426, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.375, "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": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 276.3125, "completions/mean_terminated_length": 276.3125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.0058, "grad_norm": 0.06132747605443001, "learning_rate": 1.12e-07, "loss": -0.0, "num_tokens": 328552.0, "reward": 0.31679767370224, "reward_std": 0.3372010588645935, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13359375298023224, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8249999284744263, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 835.5, "completions/mean_terminated_length": 792.0, "completions/min_length": 591.0, "completions/min_terminated_length": 591.0, "epoch": 0.006, "grad_norm": 0.0386635921895504, "learning_rate": 1.16e-07, "loss": 0.0, "num_tokens": 346496.0, "reward": 0.31500059366226196, "reward_std": 0.42705976963043213, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2549999952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.47812503576278687, "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": 0.0, "completions/max_length": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 560.875, "completions/mean_terminated_length": 560.875, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.0062, "grad_norm": 0.059056226164102554, "learning_rate": 1.2e-07, "loss": 0.0, "num_tokens": 359166.0, "reward": 0.21867355704307556, "reward_std": 0.44580790400505066, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12484374642372131, "rewards/confidence_one_or_zero": 0.8125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8681249618530273, "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": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 451.9375, "completions/mean_terminated_length": 451.9375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.0064, "grad_norm": 0.056177377700805664, "learning_rate": 1.24e-07, "loss": -0.0, "num_tokens": 370093.0, "reward": 0.5878918170928955, "reward_std": 0.5079099535942078, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23828125, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8843749761581421, "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": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 411.0625, "completions/mean_terminated_length": 411.0625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.0066, "grad_norm": 0.03778766468167305, "learning_rate": 1.28e-07, "loss": 0.0, "num_tokens": 380126.0, "reward": 0.302579402923584, "reward_std": 0.27604690194129944, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.042656250298023224, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.965624988079071, "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": 751.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 491.5625, "completions/mean_terminated_length": 491.5625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0068, "grad_norm": 0.03389393910765648, "learning_rate": 1.32e-07, "loss": 0.0, "num_tokens": 391591.0, "reward": 0.22913554310798645, "reward_std": 0.2687548100948334, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.020768748596310616, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9731249809265137, "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": 999.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 527.6875, "completions/mean_terminated_length": 527.6875, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.007, "grad_norm": 0.0393250398337841, "learning_rate": 1.36e-07, "loss": 0.0, "num_tokens": 404098.0, "reward": 0.3967946767807007, "reward_std": 0.34095028042793274, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16858750581741333, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7950000166893005, "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": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 550.9375, "completions/mean_terminated_length": 550.9375, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.0072, "grad_norm": 0.05277220904827118, "learning_rate": 1.4e-07, "loss": 0.0, "num_tokens": 416593.0, "reward": 0.25428280234336853, "reward_std": 0.4112277328968048, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07106249779462814, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9893749952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 578.4375, "completions/mean_terminated_length": 548.7333374023438, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.0074, "grad_norm": 0.054722029715776443, "learning_rate": 1.44e-07, "loss": -0.0, "num_tokens": 429656.0, "reward": 0.32523247599601746, "reward_std": 0.4097478985786438, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08796250075101852, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9087499976158142, "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": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 529.9375, "completions/mean_terminated_length": 529.9375, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.0076, "grad_norm": 0.03193988278508186, "learning_rate": 1.4799999999999998e-07, "loss": 0.0, "num_tokens": 442375.0, "reward": 0.16234511137008667, "reward_std": 0.249073788523674, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8500000238418579, "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": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 524.625, "completions/mean_terminated_length": 524.625, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.0078, "grad_norm": 0.0579211451113224, "learning_rate": 1.5199999999999998e-07, "loss": -0.0, "num_tokens": 454753.0, "reward": 0.3153136372566223, "reward_std": 0.6027231216430664, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19312500953674316, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9018750190734863, "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": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 561.8125, "completions/mean_terminated_length": 561.8125, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.008, "grad_norm": 0.020576514303684235, "learning_rate": 1.56e-07, "loss": 0.0, "num_tokens": 467694.0, "reward": 0.09984493255615234, "reward_std": 0.2149035483598709, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.918749988079071, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 815.9375, "completions/mean_terminated_length": 746.5833740234375, "completions/min_length": 452.0, "completions/min_terminated_length": 452.0, "epoch": 0.0082, "grad_norm": 0.042489923536777496, "learning_rate": 1.6e-07, "loss": 0.0, "num_tokens": 484957.0, "reward": 0.22546935081481934, "reward_std": 0.4442616403102875, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13843750953674316, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5843750238418579, "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": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 475.8125, "completions/mean_terminated_length": 475.8125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.0084, "grad_norm": 0.03728758171200752, "learning_rate": 1.64e-07, "loss": 0.0, "num_tokens": 496330.0, "reward": 0.16539205610752106, "reward_std": 0.2537340521812439, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.734375, "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": 995.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 632.625, "completions/mean_terminated_length": 632.625, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.0086, "grad_norm": 0.056957680732011795, "learning_rate": 1.68e-07, "loss": -0.0, "num_tokens": 513252.0, "reward": 0.3139849305152893, "reward_std": 0.4333817958831787, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25296875834465027, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.38750001788139343, "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": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 356.125, "completions/mean_terminated_length": 356.125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.0088, "grad_norm": 0.07416931539773941, "learning_rate": 1.7199999999999998e-07, "loss": 0.0, "num_tokens": 522742.0, "reward": 0.48375120759010315, "reward_std": 0.5627408623695374, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2175000011920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9649999737739563, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 564.5, "completions/mean_terminated_length": 533.86669921875, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.009, "grad_norm": 0.031604520976543427, "learning_rate": 1.76e-07, "loss": 0.0, "num_tokens": 535710.0, "reward": 0.2400791496038437, "reward_std": 0.2811506390571594, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.042656250298023224, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8343750238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 355.3125, "completions/mean_terminated_length": 310.73333740234375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.0092, "grad_norm": 0.0523931086063385, "learning_rate": 1.8e-07, "loss": -0.0, "num_tokens": 545747.0, "reward": 0.19421696662902832, "reward_std": 0.2593308985233307, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.013431249186396599, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9018750190734863, "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": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 445.875, "completions/mean_terminated_length": 445.875, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.0094, "grad_norm": 0.03648202493786812, "learning_rate": 1.8399999999999998e-07, "loss": 0.0, "num_tokens": 556673.0, "reward": 0.5437508821487427, "reward_std": 0.2505510151386261, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21249999105930328, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8125, "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": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 464.125, "completions/mean_terminated_length": 464.125, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.0096, "grad_norm": 0.02322164550423622, "learning_rate": 1.88e-07, "loss": -0.0, "num_tokens": 568099.0, "reward": 0.09803559631109238, "reward_std": 0.21095766127109528, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.008568749763071537, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9668124914169312, "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": 1005.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 598.125, "completions/mean_terminated_length": 598.125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.0098, "grad_norm": 0.05202183499932289, "learning_rate": 1.92e-07, "loss": -0.0, "num_tokens": 582965.0, "reward": 0.2773444950580597, "reward_std": 0.38949868083000183, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1796875, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.6781250238418579, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 459.4375, "completions/mean_terminated_length": 378.7857360839844, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.01, "grad_norm": 0.04178401827812195, "learning_rate": 1.96e-07, "loss": 0.0, "num_tokens": 593980.0, "reward": 0.17711031436920166, "reward_std": 0.2721058428287506, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.041718751192092896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8218749761581421, "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": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 356.3125, "completions/mean_terminated_length": 356.3125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0102, "grad_norm": 0.10289224982261658, "learning_rate": 2e-07, "loss": -0.0, "num_tokens": 603809.0, "reward": 0.8434168100357056, "reward_std": 0.7682167291641235, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5618312358856201, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 762.9375, "completions/mean_terminated_length": 745.5333862304688, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.0104, "grad_norm": 0.03351246565580368, "learning_rate": 2.0399999999999997e-07, "loss": 0.0, "num_tokens": 620080.0, "reward": 0.3023448586463928, "reward_std": 0.4197872579097748, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10468749701976776, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 466.0, "completions/max_terminated_length": 466.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.0106, "grad_norm": 0.08321870863437653, "learning_rate": 2.0799999999999998e-07, "loss": 0.0, "num_tokens": 629314.0, "reward": 0.19413846731185913, "reward_std": 0.40869906544685364, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07577499747276306, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.90625, "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": 586.0, "completions/max_terminated_length": 586.0, "completions/mean_length": 374.3125, "completions/mean_terminated_length": 374.3125, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.0108, "grad_norm": 0.09574051201343536, "learning_rate": 2.12e-07, "loss": 0.0, "num_tokens": 638983.0, "reward": 0.6364824771881104, "reward_std": 0.6684893369674683, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3979625105857849, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9118750095367432, "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": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 339.625, "completions/mean_terminated_length": 339.625, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.011, "grad_norm": 0.0672484040260315, "learning_rate": 2.1599999999999998e-07, "loss": 0.0, "num_tokens": 648225.0, "reward": 0.8239700198173523, "reward_std": 0.5670472383499146, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39793747663497925, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 727.25, "completions/mean_terminated_length": 658.7692260742188, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.0112, "grad_norm": 0.033097296953201294, "learning_rate": 2.1999999999999998e-07, "loss": 0.0, "num_tokens": 663877.0, "reward": 0.18109464645385742, "reward_std": 0.32951050996780396, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11218749731779099, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.703125, "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": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 444.75, "completions/mean_terminated_length": 444.75, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.0114, "grad_norm": 0.09740294516086578, "learning_rate": 2.24e-07, "loss": -0.0, "num_tokens": 675089.0, "reward": 0.5453134775161743, "reward_std": 0.6794325709342957, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34062498807907104, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.625, "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": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 358.0, "completions/mean_terminated_length": 358.0, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.0116, "grad_norm": 0.08506783843040466, "learning_rate": 2.28e-07, "loss": -0.0, "num_tokens": 684897.0, "reward": 0.5005356669425964, "reward_std": 0.6323129534721375, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.25106877088546753, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.987500011920929, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 573.4375, "completions/mean_terminated_length": 543.4000244140625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.0118, "grad_norm": 0.061930522322654724, "learning_rate": 2.32e-07, "loss": 0.0, "num_tokens": 697928.0, "reward": 0.41764459013938904, "reward_std": 0.5145143270492554, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21028751134872437, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7143750190734863, "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": 998.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 494.375, "completions/mean_terminated_length": 494.375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.012, "grad_norm": 0.03088594600558281, "learning_rate": 2.3599999999999997e-07, "loss": 0.0, "num_tokens": 710606.0, "reward": 0.06554815918207169, "reward_std": 0.17932923138141632, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9649999737739563, "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": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 255.1875, "completions/mean_terminated_length": 255.1875, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.0122, "grad_norm": 0.17090961337089539, "learning_rate": 2.4e-07, "loss": -0.0, "num_tokens": 718513.0, "reward": 1.0302009582519531, "reward_std": 0.7173442840576172, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6854000091552734, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 824.1875, "completions/mean_terminated_length": 757.5833740234375, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.0124, "grad_norm": 0.015132562257349491, "learning_rate": 2.4399999999999996e-07, "loss": 0.0, "num_tokens": 736900.0, "reward": 0.06678507477045059, "reward_std": 0.18256627023220062, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.008568749763071537, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6362500190734863, "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": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 313.875, "completions/mean_terminated_length": 313.875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.0126, "grad_norm": 0.10255121439695358, "learning_rate": 2.48e-07, "loss": 0.0, "num_tokens": 745602.0, "reward": 0.4374920129776001, "reward_std": 0.6550918817520142, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.24998125433921814, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9887499809265137, "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": 406.0, "completions/max_terminated_length": 406.0, "completions/mean_length": 293.4375, "completions/mean_terminated_length": 293.4375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.0128, "grad_norm": 0.11684675514698029, "learning_rate": 2.52e-07, "loss": -0.0, "num_tokens": 753625.0, "reward": 0.5210166573524475, "reward_std": 0.5482606291770935, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22953124344348907, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 428.125, "completions/mean_terminated_length": 428.125, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.013, "grad_norm": 0.04135340824723244, "learning_rate": 2.56e-07, "loss": 0.0, "num_tokens": 764219.0, "reward": 0.09375117719173431, "reward_std": 0.27195286750793457, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0625, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.7418749928474426, "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": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 458.25, "completions/mean_terminated_length": 458.25, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.0132, "grad_norm": 0.08070624619722366, "learning_rate": 2.6e-07, "loss": 0.0, "num_tokens": 775247.0, "reward": 0.3131203055381775, "reward_std": 0.602284848690033, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.18873749673366547, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9856250286102295, "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": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 552.25, "completions/mean_terminated_length": 552.25, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.0134, "grad_norm": 0.05386007949709892, "learning_rate": 2.64e-07, "loss": 0.0, "num_tokens": 788691.0, "reward": 0.2915632426738739, "reward_std": 0.40394195914268494, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20812499523162842, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.6237499713897705, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 793.6875, "completions/mean_terminated_length": 716.9166870117188, "completions/min_length": 419.0, "completions/min_terminated_length": 419.0, "epoch": 0.0136, "grad_norm": 0.018275510519742966, "learning_rate": 2.68e-07, "loss": -0.0, "num_tokens": 805342.0, "reward": 0.04718823730945587, "reward_std": 0.1887499839067459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03187499940395355, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.48750001192092896, "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": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 354.0625, "completions/mean_terminated_length": 354.0625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.0138, "grad_norm": 0.09318700432777405, "learning_rate": 2.72e-07, "loss": -0.0, "num_tokens": 814559.0, "reward": 0.31841984391212463, "reward_std": 0.6044764518737793, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19933749735355377, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 444.875, "completions/mean_terminated_length": 444.875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.014, "grad_norm": 0.03874102607369423, "learning_rate": 2.7600000000000004e-07, "loss": -0.0, "num_tokens": 825677.0, "reward": 0.3620040714740753, "reward_std": 0.4046049118041992, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09900625795125961, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 0.0, "completions/max_length": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 431.75, "completions/mean_terminated_length": 431.75, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0142, "grad_norm": 0.0679880678653717, "learning_rate": 2.8e-07, "loss": -0.0, "num_tokens": 836553.0, "reward": 0.3557697534561157, "reward_std": 0.5434787273406982, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21153749525547028, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7875000238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 528.125, "completions/mean_terminated_length": 495.0666809082031, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.0144, "grad_norm": 0.062334757298231125, "learning_rate": 2.8399999999999995e-07, "loss": -0.0, "num_tokens": 849355.0, "reward": 0.281147837638855, "reward_std": 0.6044476628303528, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.18729375302791595, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9181250333786011, "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": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 272.3125, "completions/mean_terminated_length": 272.3125, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 0.0146, "grad_norm": 0.12119140475988388, "learning_rate": 2.88e-07, "loss": 0.0, "num_tokens": 857408.0, "reward": 0.6607667207717896, "reward_std": 0.698896586894989, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3840312659740448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 524.9375, "completions/mean_terminated_length": 524.9375, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.0148, "grad_norm": 0.03456779569387436, "learning_rate": 2.9199999999999997e-07, "loss": -0.0, "num_tokens": 871247.0, "reward": 0.3625760078430176, "reward_std": 0.4049436151981354, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10014999657869339, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 841.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 627.375, "completions/mean_terminated_length": 627.375, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.015, "grad_norm": 0.04225312918424606, "learning_rate": 2.9599999999999995e-07, "loss": -0.0, "num_tokens": 885669.0, "reward": 0.1873447149991989, "reward_std": 0.5119205713272095, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12468750029802322, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.778124988079071, "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": 0.0, "completions/max_length": 477.0, "completions/max_terminated_length": 477.0, "completions/mean_length": 384.0625, "completions/mean_terminated_length": 384.0625, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.0152, "grad_norm": 0.038098715245723724, "learning_rate": 3e-07, "loss": -0.0, "num_tokens": 895942.0, "reward": 0.20031052827835083, "reward_std": 0.2674039900302887, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.025618748739361763, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9724999666213989, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 764.125, "completions/mean_terminated_length": 677.5, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.0154, "grad_norm": 3.726798070147197e-08, "learning_rate": 3.0399999999999997e-07, "loss": -0.0, "num_tokens": 914088.0, "reward": 4.899999908047903e-07, "reward_std": 4.32512052839229e-07, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.0, "rewards/mean_confidence_reward": 0.49000000953674316, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 665.5, "completions/mean_terminated_length": 641.6000366210938, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.0156, "grad_norm": 0.04155049845576286, "learning_rate": 3.08e-07, "loss": 0.0, "num_tokens": 929504.0, "reward": 0.09374824166297913, "reward_std": 0.37498739361763, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06249374896287918, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.9275000095367432, "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": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 244.875, "completions/mean_terminated_length": 244.875, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.0158, "grad_norm": 0.08507490158081055, "learning_rate": 3.12e-07, "loss": -0.0, "num_tokens": 937102.0, "reward": 0.39749467372894287, "reward_std": 0.3954924941062927, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10748749226331711, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 664.4375, "completions/mean_terminated_length": 664.4375, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.016, "grad_norm": 0.045550789684057236, "learning_rate": 3.1599999999999997e-07, "loss": 0.0, "num_tokens": 952565.0, "reward": 0.3835945427417755, "reward_std": 0.41363781690597534, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2671874761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7281249761581421, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 622.5, "completions/mean_terminated_length": 440.0, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.0162, "grad_norm": 0.033813316375017166, "learning_rate": 3.2e-07, "loss": 0.0, "num_tokens": 966829.0, "reward": 0.1400790512561798, "reward_std": 0.2507926821708679, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.030156249180436134, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.6643750071525574, "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": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 560.4375, "completions/mean_terminated_length": 560.4375, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.0164, "grad_norm": 0.04374932870268822, "learning_rate": 3.24e-07, "loss": 0.0, "num_tokens": 980916.0, "reward": 0.1960698962211609, "reward_std": 0.3191184401512146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07963749766349792, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8831250667572021, "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": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 255.0, "completions/mean_terminated_length": 255.0, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.0166, "grad_norm": 0.07968038320541382, "learning_rate": 3.28e-07, "loss": 0.0, "num_tokens": 988852.0, "reward": 0.5718728303909302, "reward_std": 0.4655292332172394, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.26874375343322754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8149999380111694, "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": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 533.375, "completions/mean_terminated_length": 533.375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.0168, "grad_norm": 0.054786667227745056, "learning_rate": 3.32e-07, "loss": 0.0, "num_tokens": 1001594.0, "reward": 0.16187529265880585, "reward_std": 0.35821738839149475, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13625000417232513, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.16875000298023224, "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": 591.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 408.5625, "completions/mean_terminated_length": 408.5625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.017, "grad_norm": 0.08940014243125916, "learning_rate": 3.36e-07, "loss": -0.0, "num_tokens": 1012499.0, "reward": 0.48078250885009766, "reward_std": 0.5626484751701355, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21156249940395355, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9081250429153442, "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": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 201.0625, "completions/mean_terminated_length": 201.0625, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.0172, "grad_norm": 0.030316654592752457, "learning_rate": 3.4000000000000003e-07, "loss": 0.0, "num_tokens": 1019588.0, "reward": 0.445523202419281, "reward_std": 0.17429538071155548, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01604374870657921, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9906250238418579, "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": 0.0, "completions/max_length": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 462.3125, "completions/mean_terminated_length": 462.3125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.0174, "grad_norm": 0.08344651758670807, "learning_rate": 3.4399999999999996e-07, "loss": -0.0, "num_tokens": 1031001.0, "reward": 0.4372665286064148, "reward_std": 0.6796832084655762, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31203123927116394, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.6806249618530273, "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": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 472.5625, "completions/mean_terminated_length": 472.5625, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.0176, "grad_norm": 0.03455330803990364, "learning_rate": 3.4799999999999994e-07, "loss": 0.0, "num_tokens": 1042194.0, "reward": 0.22856354713439941, "reward_std": 0.2679406404495239, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.019625000655651093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 688.0, "completions/max_terminated_length": 688.0, "completions/mean_length": 435.6875, "completions/mean_terminated_length": 435.6875, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.0178, "grad_norm": 0.09790806472301483, "learning_rate": 3.52e-07, "loss": 0.0, "num_tokens": 1052877.0, "reward": 0.37498903274536133, "reward_std": 0.670798122882843, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.24997499585151672, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 504.875, "completions/mean_terminated_length": 504.875, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.018, "grad_norm": 0.033641744405031204, "learning_rate": 3.5599999999999996e-07, "loss": -0.0, "num_tokens": 1065531.0, "reward": 0.16601359844207764, "reward_std": 0.2546195387840271, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01952499896287918, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9112499952316284, "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": 500.0, "completions/max_terminated_length": 500.0, "completions/mean_length": 371.875, "completions/mean_terminated_length": 371.875, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.0182, "grad_norm": 0.0929093137383461, "learning_rate": 3.6e-07, "loss": 0.0, "num_tokens": 1075801.0, "reward": 0.6931260228157043, "reward_std": 0.7020448446273804, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44875001907348633, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8925000429153442, "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": 787.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 510.4375, "completions/mean_terminated_length": 510.4375, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.0184, "grad_norm": 5.657363644218094e-08, "learning_rate": 3.64e-07, "loss": 0.0, "num_tokens": 1088192.0, "reward": 1.1012500635843026e-06, "reward_std": 5.067659003543667e-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": 0.9137499928474426, "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": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 587.75, "completions/mean_terminated_length": 587.75, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.0186, "grad_norm": 0.027916966006159782, "learning_rate": 3.6799999999999996e-07, "loss": 0.0, "num_tokens": 1102348.0, "reward": 0.1966918706893921, "reward_std": 0.2625122666358948, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01838124915957451, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9731250405311584, "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": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 379.9375, "completions/mean_terminated_length": 379.9375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.0188, "grad_norm": 0.0944044217467308, "learning_rate": 3.72e-07, "loss": 0.0, "num_tokens": 1112459.0, "reward": 0.5028729438781738, "reward_std": 0.7069486975669861, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31824374198913574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9775000214576721, "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": 853.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 470.125, "completions/mean_terminated_length": 470.125, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.019, "grad_norm": 0.08061134070158005, "learning_rate": 3.76e-07, "loss": -0.0, "num_tokens": 1123805.0, "reward": 0.3865637183189392, "reward_std": 0.5940209031105042, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21062499284744263, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.921875, "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": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 425.5625, "completions/mean_terminated_length": 425.5625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.0192, "grad_norm": 0.08343944698572159, "learning_rate": 3.7999999999999996e-07, "loss": 0.0, "num_tokens": 1134230.0, "reward": 0.7125789523124695, "reward_std": 0.553808331489563, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.42515623569488525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8031250238418579, "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": 178.0, "completions/max_terminated_length": 178.0, "completions/mean_length": 141.375, "completions/mean_terminated_length": 141.375, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "epoch": 0.0194, "grad_norm": 0.19533324241638184, "learning_rate": 3.84e-07, "loss": -0.0, "num_tokens": 1140252.0, "reward": 0.9685763120651245, "reward_std": 0.7179321050643921, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6246500015258789, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9862499833106995, "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": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 428.25, "completions/mean_terminated_length": 428.25, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.0196, "grad_norm": 0.08816035836935043, "learning_rate": 3.88e-07, "loss": -0.0, "num_tokens": 1150576.0, "reward": 0.489610493183136, "reward_std": 0.5638236403465271, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2292187511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 394.0, "completions/max_terminated_length": 394.0, "completions/mean_length": 318.625, "completions/mean_terminated_length": 318.625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0198, "grad_norm": 0.09611859172582626, "learning_rate": 3.92e-07, "loss": -0.0, "num_tokens": 1160058.0, "reward": 0.5775010585784912, "reward_std": 0.6004603505134583, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2800000011920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9606250524520874, "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": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.02, "grad_norm": 0.038361892104148865, "learning_rate": 3.96e-07, "loss": -0.0, "num_tokens": 1169308.0, "reward": 0.44281357526779175, "reward_std": 0.22002898156642914, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07312499731779099, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 282.8125, "completions/mean_terminated_length": 282.8125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.0202, "grad_norm": 0.08380813151597977, "learning_rate": 4e-07, "loss": -0.0, "num_tokens": 1178009.0, "reward": 0.40975117683410645, "reward_std": 0.5840522646903992, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19449999928474426, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8425000309944153, "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": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 402.9375, "completions/mean_terminated_length": 402.9375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.0204, "grad_norm": 0.050619881600141525, "learning_rate": 4.04e-07, "loss": -0.0, "num_tokens": 1188488.0, "reward": 0.19109177589416504, "reward_std": 0.40605229139328003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06968124955892563, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9662500619888306, "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": 356.0, "completions/max_terminated_length": 356.0, "completions/mean_length": 256.9375, "completions/mean_terminated_length": 256.9375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.0206, "grad_norm": 0.08154933154582977, "learning_rate": 4.0799999999999995e-07, "loss": 0.0, "num_tokens": 1196503.0, "reward": 0.517013669013977, "reward_std": 0.4834006130695343, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22152501344680786, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8474999666213989, "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": 442.0, "completions/max_terminated_length": 442.0, "completions/mean_length": 348.3125, "completions/mean_terminated_length": 348.3125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.0208, "grad_norm": 0.07601585984230042, "learning_rate": 4.12e-07, "loss": 0.0, "num_tokens": 1205740.0, "reward": 0.9552228450775146, "reward_std": 0.5876926183700562, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5354437232017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 186.625, "completions/mean_terminated_length": 186.625, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.021, "grad_norm": 0.08451579511165619, "learning_rate": 4.1599999999999997e-07, "loss": 0.0, "num_tokens": 1212742.0, "reward": 0.5422293543815613, "reward_std": 0.4262162744998932, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14695626497268677, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9812500476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 439.6875, "completions/mean_terminated_length": 400.7333679199219, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.0212, "grad_norm": 0.03513219580054283, "learning_rate": 4.1999999999999995e-07, "loss": 0.0, "num_tokens": 1224017.0, "reward": 0.30304503440856934, "reward_std": 0.2769280672073364, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04358749836683273, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9056249856948853, "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": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 366.25, "completions/mean_terminated_length": 366.25, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.0214, "grad_norm": 0.0607839897274971, "learning_rate": 4.24e-07, "loss": 0.0, "num_tokens": 1234005.0, "reward": 0.2862507104873657, "reward_std": 0.43953564763069153, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25999999046325684, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.453125, "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": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 375.4375, "completions/mean_terminated_length": 375.4375, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.0216, "grad_norm": 0.02885805070400238, "learning_rate": 4.2799999999999997e-07, "loss": 0.0, "num_tokens": 1244300.0, "reward": 0.10289157927036285, "reward_std": 0.22120818495750427, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9556249976158142, "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": 259.0, "completions/max_terminated_length": 259.0, "completions/mean_length": 192.75, "completions/mean_terminated_length": 192.75, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.0218, "grad_norm": 0.17998896539211273, "learning_rate": 4.3199999999999995e-07, "loss": -0.0, "num_tokens": 1251320.0, "reward": 0.9119386672973633, "reward_std": 0.7083633542060852, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5738750100135803, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9718124866485596, "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": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 328.3125, "completions/mean_terminated_length": 328.3125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.022, "grad_norm": 0.04378411918878555, "learning_rate": 4.36e-07, "loss": 0.0, "num_tokens": 1260573.0, "reward": 0.23388537764549255, "reward_std": 0.27463752031326294, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.030268749222159386, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9312499761581421, "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": 582.0, "completions/max_terminated_length": 582.0, "completions/mean_length": 426.375, "completions/mean_terminated_length": 426.375, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0222, "grad_norm": 0.030276762321591377, "learning_rate": 4.3999999999999997e-07, "loss": -0.0, "num_tokens": 1271059.0, "reward": 0.13233865797519684, "reward_std": 0.2369442731142044, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.014674998819828033, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 491.0, "completions/mean_terminated_length": 491.0, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.0224, "grad_norm": 0.06538961082696915, "learning_rate": 4.44e-07, "loss": -0.0, "num_tokens": 1282563.0, "reward": 0.41390711069107056, "reward_std": 0.4795403480529785, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2653124928474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8287500143051147, "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": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 266.0625, "completions/mean_terminated_length": 266.0625, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0226, "grad_norm": 0.10549194365739822, "learning_rate": 4.48e-07, "loss": 0.0, "num_tokens": 1290644.0, "reward": 0.9125732779502869, "reward_std": 0.6334460377693176, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5126437544822693, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9143750071525574, "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": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 422.8125, "completions/mean_terminated_length": 422.8125, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.0228, "grad_norm": 0.08699935674667358, "learning_rate": 4.5199999999999997e-07, "loss": -0.0, "num_tokens": 1303729.0, "reward": 0.7243759632110596, "reward_std": 0.7272898554801941, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.44875001907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 339.0, "completions/max_terminated_length": 339.0, "completions/mean_length": 228.5625, "completions/mean_terminated_length": 228.5625, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.023, "grad_norm": 0.12602975964546204, "learning_rate": 4.56e-07, "loss": 0.0, "num_tokens": 1311386.0, "reward": 0.8773416876792908, "reward_std": 0.6692045331001282, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5046812295913696, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 482.0, "completions/mean_terminated_length": 482.0, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0232, "grad_norm": 0.06035184487700462, "learning_rate": 4.6e-07, "loss": -0.0, "num_tokens": 1323098.0, "reward": 0.39929765462875366, "reward_std": 0.48166120052337646, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.36109375953674316, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.4749999940395355, "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": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 451.1875, "completions/mean_terminated_length": 451.1875, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.0234, "grad_norm": 0.033002082258462906, "learning_rate": 4.64e-07, "loss": 0.0, "num_tokens": 1336525.0, "reward": 0.16725116968154907, "reward_std": 0.2563927471637726, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02199999988079071, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8518750071525574, "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": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 262.125, "completions/mean_terminated_length": 262.125, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.0236, "grad_norm": 0.08861807733774185, "learning_rate": 4.68e-07, "loss": 0.0, "num_tokens": 1344655.0, "reward": 0.7781258821487427, "reward_std": 0.525794267654419, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4937500059604645, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.39375001192092896, "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": 774.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 337.625, "completions/mean_terminated_length": 337.625, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0238, "grad_norm": 0.09875895828008652, "learning_rate": 4.7199999999999994e-07, "loss": 0.0, "num_tokens": 1353849.0, "reward": 0.90707927942276, "reward_std": 0.7111445069313049, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5641562342643738, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9424999952316284, "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": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 332.0, "completions/mean_terminated_length": 332.0, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.024, "grad_norm": 0.06012330576777458, "learning_rate": 4.76e-07, "loss": 0.0, "num_tokens": 1362937.0, "reward": 0.3613823354244232, "reward_std": 0.40436989068984985, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09776249527931213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 482.375, "completions/mean_terminated_length": 446.2666931152344, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.0242, "grad_norm": 0.03735286369919777, "learning_rate": 4.8e-07, "loss": 0.0, "num_tokens": 1374975.0, "reward": 0.13078220188617706, "reward_std": 0.29649245738983154, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07406249642372131, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6943750381469727, "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": 0.0, "completions/max_length": 206.0, "completions/max_terminated_length": 206.0, "completions/mean_length": 184.1875, "completions/mean_terminated_length": 184.1875, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.0244, "grad_norm": 0.16276690363883972, "learning_rate": 4.839999999999999e-07, "loss": -0.0, "num_tokens": 1381586.0, "reward": 0.9066233038902283, "reward_std": 0.7116478681564331, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5632437467575073, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9868749976158142, "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": 0.0, "completions/max_length": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 429.0, "completions/mean_terminated_length": 429.0, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.0246, "grad_norm": 0.042794376611709595, "learning_rate": 4.879999999999999e-07, "loss": 0.0, "num_tokens": 1392770.0, "reward": 0.24007928371429443, "reward_std": 0.2811505198478699, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.042656250298023224, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 638.25, "completions/mean_terminated_length": 549.2307739257812, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.0248, "grad_norm": 0.04814385995268822, "learning_rate": 4.92e-07, "loss": -0.0, "num_tokens": 1406742.0, "reward": 0.24461017549037933, "reward_std": 0.34290605783462524, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11421874165534973, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5062500238418579, "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": 0.0, "completions/max_length": 208.0, "completions/max_terminated_length": 208.0, "completions/mean_length": 176.8125, "completions/mean_terminated_length": 176.8125, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.025, "grad_norm": 0.17295581102371216, "learning_rate": 4.96e-07, "loss": -0.0, "num_tokens": 1413347.0, "reward": 1.0311481952667236, "reward_std": 0.7179985642433167, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6872937083244324, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9293749928474426, "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": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 361.3125, "completions/mean_terminated_length": 361.3125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.0252, "grad_norm": 0.10872635245323181, "learning_rate": 5e-07, "loss": 0.0, "num_tokens": 1422808.0, "reward": 0.6559324860572815, "reward_std": 0.7681480646133423, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4368624985218048, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 0.0, "completions/max_length": 824.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 444.4375, "completions/mean_terminated_length": 444.4375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.0254, "grad_norm": 0.04795417562127113, "learning_rate": 5.04e-07, "loss": -0.0, "num_tokens": 1433727.0, "reward": 0.3888257145881653, "reward_std": 0.37516841292381287, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21514999866485596, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.6806250214576721, "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": 467.0, "completions/max_terminated_length": 467.0, "completions/mean_length": 416.1875, "completions/mean_terminated_length": 416.1875, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.0256, "grad_norm": 0.021793320775032043, "learning_rate": 5.079999999999999e-07, "loss": -0.0, "num_tokens": 1444610.0, "reward": 0.06436053663492203, "reward_std": 0.17587241530418396, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.003718750085681677, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9818750023841858, "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": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 353.5, "completions/mean_terminated_length": 353.5, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.0258, "grad_norm": 0.11864348500967026, "learning_rate": 5.12e-07, "loss": -0.0, "num_tokens": 1454362.0, "reward": 0.6560139656066895, "reward_std": 0.7682431936264038, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4370250105857849, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 0.0, "completions/max_length": 523.0, "completions/max_terminated_length": 523.0, "completions/mean_length": 387.875, "completions/mean_terminated_length": 387.875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.026, "grad_norm": 0.038355495780706406, "learning_rate": 5.16e-07, "loss": 0.0, "num_tokens": 1464872.0, "reward": 0.1371885985136032, "reward_std": 0.24540835618972778, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02437499910593033, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 407.0625, "completions/mean_terminated_length": 407.0625, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.0262, "grad_norm": 0.04656800255179405, "learning_rate": 5.2e-07, "loss": 0.0, "num_tokens": 1474905.0, "reward": 0.5924010276794434, "reward_std": 0.398771196603775, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18479998409748077, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 751.0, "completions/max_terminated_length": 751.0, "completions/mean_length": 318.75, "completions/mean_terminated_length": 318.75, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.0264, "grad_norm": 0.05038650333881378, "learning_rate": 5.24e-07, "loss": -0.0, "num_tokens": 1484229.0, "reward": 0.37929511070251465, "reward_std": 0.2756575644016266, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07108750194311142, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8675000071525574, "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": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 417.9375, "completions/mean_terminated_length": 417.9375, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0266, "grad_norm": 0.02768278308212757, "learning_rate": 5.28e-07, "loss": 0.0, "num_tokens": 1495236.0, "reward": 0.099845290184021, "reward_std": 0.21490338444709778, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9818750023841858, "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": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 306.625, "completions/mean_terminated_length": 306.625, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0268, "grad_norm": 0.07502417266368866, "learning_rate": 5.32e-07, "loss": 0.0, "num_tokens": 1503982.0, "reward": 0.368047833442688, "reward_std": 0.4072614014148712, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11109375208616257, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8924999833106995, "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": 708.0, "completions/max_terminated_length": 708.0, "completions/mean_length": 400.3125, "completions/mean_terminated_length": 400.3125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.027, "grad_norm": 0.06546711176633835, "learning_rate": 5.36e-07, "loss": -0.0, "num_tokens": 1513971.0, "reward": 0.42140424251556396, "reward_std": 0.49399352073669434, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15530623495578766, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9568750262260437, "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": 0.0, "completions/max_length": 899.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 517.25, "completions/mean_terminated_length": 517.25, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.0272, "grad_norm": 0.09113264083862305, "learning_rate": 5.4e-07, "loss": 0.0, "num_tokens": 1525943.0, "reward": 0.6265323162078857, "reward_std": 0.7180452346801758, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.37806248664855957, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9668749570846558, "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": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 356.0, "completions/mean_terminated_length": 356.0, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0274, "grad_norm": 0.05682982876896858, "learning_rate": 5.44e-07, "loss": -0.0, "num_tokens": 1536087.0, "reward": 0.4256134331226349, "reward_std": 0.49471327662467957, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16372498869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8943749666213989, "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": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 522.4375, "completions/mean_terminated_length": 522.4375, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.0276, "grad_norm": 0.06273636966943741, "learning_rate": 5.48e-07, "loss": -0.0, "num_tokens": 1548446.0, "reward": 0.46866992115974426, "reward_std": 0.6445512771606445, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3748374879360199, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.4962500035762787, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 779.0, "completions/mean_terminated_length": 762.6666870117188, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.0278, "grad_norm": 0.0419401191174984, "learning_rate": 5.520000000000001e-07, "loss": 0.0, "num_tokens": 1564910.0, "reward": 0.19093841314315796, "reward_std": 0.3597692847251892, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13187499344348907, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.6612499952316284, "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": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 310.625, "completions/mean_terminated_length": 310.625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.028, "grad_norm": 0.11100335419178009, "learning_rate": 5.560000000000001e-07, "loss": -0.0, "num_tokens": 1573672.0, "reward": 0.9370856881141663, "reward_std": 0.7496675252914429, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6241687536239624, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9768750667572021, "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": 0.0, "completions/max_length": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 544.3125, "completions/mean_terminated_length": 544.3125, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.0282, "grad_norm": 0.052290357649326324, "learning_rate": 5.6e-07, "loss": -0.0, "num_tokens": 1586141.0, "reward": 0.44647321105003357, "reward_std": 0.5747976303100586, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20544375479221344, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9724999666213989, "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": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 698.125, "completions/mean_terminated_length": 698.125, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.0284, "grad_norm": 0.057536885142326355, "learning_rate": 5.639999999999999e-07, "loss": 0.0, "num_tokens": 1601759.0, "reward": 0.23375047743320465, "reward_std": 0.41852709650993347, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2175000011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.484375, "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": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 377.0625, "completions/mean_terminated_length": 377.0625, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0286, "grad_norm": 0.06383604556322098, "learning_rate": 5.679999999999999e-07, "loss": 0.0, "num_tokens": 1612240.0, "reward": 0.18742303550243378, "reward_std": 0.5121341943740845, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12484374642372131, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 0.0, "completions/max_length": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 367.6875, "completions/mean_terminated_length": 367.6875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.0288, "grad_norm": 0.09325961023569107, "learning_rate": 5.719999999999999e-07, "loss": -0.0, "num_tokens": 1621851.0, "reward": 0.4685760736465454, "reward_std": 0.7178022265434265, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3121500015258789, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9762500524520874, "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": 338.0, "completions/max_terminated_length": 338.0, "completions/mean_length": 275.875, "completions/mean_terminated_length": 275.875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.029, "grad_norm": 0.09880378097295761, "learning_rate": 5.76e-07, "loss": -0.0, "num_tokens": 1629705.0, "reward": 0.7458227872848511, "reward_std": 0.5658391714096069, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3666437268257141, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 442.75, "completions/mean_terminated_length": 442.75, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.0292, "grad_norm": 0.038493990898132324, "learning_rate": 5.8e-07, "loss": -0.0, "num_tokens": 1640773.0, "reward": 0.29343897104263306, "reward_std": 0.2679325044155121, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.02437499910593033, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9749999642372131, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 771.875, "completions/mean_terminated_length": 687.8333740234375, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.0294, "grad_norm": 0.010580530390143394, "learning_rate": 5.839999999999999e-07, "loss": -0.0, "num_tokens": 1657859.0, "reward": 0.034297917038202286, "reward_std": 0.13718746602535248, "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": 0.6687500476837158, "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": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 272.3125, "completions/mean_terminated_length": 272.3125, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 0.0296, "grad_norm": 0.09305991232395172, "learning_rate": 5.879999999999999e-07, "loss": -0.0, "num_tokens": 1666088.0, "reward": 0.9476261734962463, "reward_std": 0.5940559506416321, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5202499628067017, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9737499952316284, "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": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 558.75, "completions/mean_terminated_length": 558.75, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.0298, "grad_norm": 0.05057363584637642, "learning_rate": 5.919999999999999e-07, "loss": 0.0, "num_tokens": 1679060.0, "reward": 0.21929500997066498, "reward_std": 0.5155492424964905, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1260875016450882, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 389.9375, "completions/mean_terminated_length": 389.9375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.03, "grad_norm": 0.10617473721504211, "learning_rate": 5.96e-07, "loss": 0.0, "num_tokens": 1689411.0, "reward": 0.7840481996536255, "reward_std": 0.7507089376449585, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5055937170982361, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9162499904632568, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 653.625, "completions/mean_terminated_length": 628.933349609375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.0302, "grad_norm": 0.045913152396678925, "learning_rate": 6e-07, "loss": 0.0, "num_tokens": 1703709.0, "reward": 0.09375107288360596, "reward_std": 0.3750002384185791, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.8856250047683716, "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": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 329.4375, "completions/mean_terminated_length": 329.4375, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.0304, "grad_norm": 0.11458048224449158, "learning_rate": 6.04e-07, "loss": -0.0, "num_tokens": 1712468.0, "reward": 0.787016749382019, "reward_std": 0.6795200109481812, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4490312337875366, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 820.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 425.8125, "completions/mean_terminated_length": 425.8125, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.0306, "grad_norm": 0.06713155657052994, "learning_rate": 6.079999999999999e-07, "loss": 0.0, "num_tokens": 1722913.0, "reward": 0.48382607102394104, "reward_std": 0.5628854036331177, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21764999628067017, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9387499690055847, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 847.375, "completions/mean_terminated_length": 788.5, "completions/min_length": 551.0, "completions/min_terminated_length": 551.0, "epoch": 0.0308, "grad_norm": 0.015245222486555576, "learning_rate": 6.119999999999999e-07, "loss": 0.0, "num_tokens": 1740871.0, "reward": 0.04718823730945587, "reward_std": 0.1887499839067459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03187499940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.546875, "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": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 228.9375, "completions/mean_terminated_length": 228.9375, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.031, "grad_norm": 0.0922832265496254, "learning_rate": 6.16e-07, "loss": -0.0, "num_tokens": 1748102.0, "reward": 0.5434359908103943, "reward_std": 0.4265536665916443, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1493687480688095, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9837499856948853, "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": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 594.375, "completions/mean_terminated_length": 594.375, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.0312, "grad_norm": 0.05348094925284386, "learning_rate": 6.2e-07, "loss": 0.0, "num_tokens": 1761740.0, "reward": 0.3180353343486786, "reward_std": 0.437282532453537, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13606874644756317, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8987500071525574, "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": 314.0, "completions/max_terminated_length": 314.0, "completions/mean_length": 156.5625, "completions/mean_terminated_length": 156.5625, "completions/min_length": 100.0, "completions/min_terminated_length": 100.0, "epoch": 0.0314, "grad_norm": 0.07742555439472198, "learning_rate": 6.24e-07, "loss": 0.0, "num_tokens": 1767573.0, "reward": 0.3823951184749603, "reward_std": 0.2283220738172531, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.014787498861551285, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9237499833106995, "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": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 448.8125, "completions/mean_terminated_length": 448.8125, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.0316, "grad_norm": 0.043902650475502014, "learning_rate": 6.28e-07, "loss": 0.0, "num_tokens": 1778226.0, "reward": 0.3007700443267822, "reward_std": 0.27441585063934326, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03903749957680702, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9737499952316284, "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": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 232.6875, "completions/mean_terminated_length": 232.6875, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.0318, "grad_norm": 0.16233301162719727, "learning_rate": 6.319999999999999e-07, "loss": -0.0, "num_tokens": 1785437.0, "reward": 0.8232478499412537, "reward_std": 0.6494977474212646, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45899373292922974, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 257.0, "completions/max_terminated_length": 257.0, "completions/mean_length": 157.5625, "completions/mean_terminated_length": 157.5625, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.032, "grad_norm": 0.16797572374343872, "learning_rate": 6.36e-07, "loss": -0.0, "num_tokens": 1791686.0, "reward": 1.0933387279510498, "reward_std": 0.6635278463363647, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.749174952507019, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8462499976158142, "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": 0.0, "completions/max_length": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 311.1875, "completions/mean_terminated_length": 311.1875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.0322, "grad_norm": 0.07733765989542007, "learning_rate": 6.4e-07, "loss": -0.0, "num_tokens": 1800617.0, "reward": 0.543879508972168, "reward_std": 0.5311792492866516, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.21275624632835388, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8562500476837158, "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": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 266.875, "completions/mean_terminated_length": 266.875, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.0324, "grad_norm": 0.1297840178012848, "learning_rate": 6.44e-07, "loss": -0.0, "num_tokens": 1808759.0, "reward": 0.7186353206634521, "reward_std": 0.7519524097442627, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49976876378059387, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9137499928474426, "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": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 272.625, "completions/mean_terminated_length": 272.625, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 0.0326, "grad_norm": 0.12264753878116608, "learning_rate": 6.48e-07, "loss": -0.0, "num_tokens": 1816993.0, "reward": 0.6903858184814453, "reward_std": 0.6790797114372253, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.380768746137619, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 544.0, "completions/max_terminated_length": 544.0, "completions/mean_length": 423.0625, "completions/mean_terminated_length": 423.0625, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.0328, "grad_norm": 0.057838261127471924, "learning_rate": 6.52e-07, "loss": 0.0, "num_tokens": 1827810.0, "reward": 0.8708571791648865, "reward_std": 0.48008251190185547, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4917125105857849, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8681249618530273, "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": 1011.0, "completions/mean_length": 817.8125, "completions/mean_terminated_length": 804.0667114257812, "completions/min_length": 549.0, "completions/min_terminated_length": 549.0, "epoch": 0.033, "grad_norm": 0.042156293988227844, "learning_rate": 6.56e-07, "loss": 0.0, "num_tokens": 1845359.0, "reward": 0.20836006104946136, "reward_std": 0.3826383948326111, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1667187511920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5718749761581421, "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": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 531.375, "completions/mean_terminated_length": 531.375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.0332, "grad_norm": 0.03636961802840233, "learning_rate": 6.6e-07, "loss": 0.0, "num_tokens": 1857909.0, "reward": 0.19726386666297913, "reward_std": 0.2634170651435852, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01952499896287918, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9137499928474426, "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": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 465.1875, "completions/mean_terminated_length": 465.1875, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.0334, "grad_norm": 0.07119625806808472, "learning_rate": 6.64e-07, "loss": 0.0, "num_tokens": 1869416.0, "reward": 0.4525009095668793, "reward_std": 0.5755164623260498, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2175000011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.84375, "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": 276.0, "completions/max_terminated_length": 276.0, "completions/mean_length": 176.125, "completions/mean_terminated_length": 176.125, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "epoch": 0.0336, "grad_norm": 0.11353052407503128, "learning_rate": 6.68e-07, "loss": -0.0, "num_tokens": 1875866.0, "reward": 1.284519910812378, "reward_std": 0.47554677724838257, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8190374970436096, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9781250357627869, "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": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 240.375, "completions/mean_terminated_length": 240.375, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.0338, "grad_norm": 0.09399046748876572, "learning_rate": 6.72e-07, "loss": 0.0, "num_tokens": 1883168.0, "reward": 0.5580259561538696, "reward_std": 0.4252946972846985, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17855000495910645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 276.0, "completions/max_terminated_length": 276.0, "completions/mean_length": 195.0, "completions/mean_terminated_length": 195.0, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.034, "grad_norm": 0.06324208527803421, "learning_rate": 6.76e-07, "loss": -0.0, "num_tokens": 1890480.0, "reward": 0.7094197273254395, "reward_std": 0.3924414813518524, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23133748769760132, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "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": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 592.0625, "completions/mean_terminated_length": 592.0625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.0342, "grad_norm": 0.058833274990320206, "learning_rate": 6.800000000000001e-07, "loss": -0.0, "num_tokens": 1903777.0, "reward": 0.3153136372566223, "reward_std": 0.6027231216430664, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19312499463558197, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9612499475479126, "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": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 396.8125, "completions/mean_terminated_length": 396.8125, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.0344, "grad_norm": 0.05650670453906059, "learning_rate": 6.84e-07, "loss": 0.0, "num_tokens": 1914350.0, "reward": 0.2878890037536621, "reward_std": 0.4108920991420746, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.07577500492334366, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 174.3125, "completions/mean_terminated_length": 174.3125, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 0.0346, "grad_norm": 0.15495720505714417, "learning_rate": 6.879999999999999e-07, "loss": -0.0, "num_tokens": 1920867.0, "reward": 0.8513918519020081, "reward_std": 0.6201745867729187, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4527812600135803, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9774999618530273, "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": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 376.5625, "completions/mean_terminated_length": 376.5625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0348, "grad_norm": 0.06440185755491257, "learning_rate": 6.919999999999999e-07, "loss": 0.0, "num_tokens": 1930572.0, "reward": 0.297644704580307, "reward_std": 0.41670194268226624, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0952875018119812, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.953125, "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": 0.0, "completions/max_length": 848.0, "completions/max_terminated_length": 848.0, "completions/mean_length": 486.8125, "completions/mean_terminated_length": 486.8125, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.035, "grad_norm": 0.0538332499563694, "learning_rate": 6.959999999999999e-07, "loss": 0.0, "num_tokens": 1942937.0, "reward": 0.19663846492767334, "reward_std": 0.41106116771698, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08077499270439148, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8360625505447388, "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": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 520.125, "completions/mean_terminated_length": 520.125, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.0352, "grad_norm": 0.08643592894077301, "learning_rate": 7e-07, "loss": -0.0, "num_tokens": 1955435.0, "reward": 0.28101643919944763, "reward_std": 0.6041654944419861, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.18703125417232513, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6437499523162842, "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": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 509.375, "completions/mean_terminated_length": 509.375, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.0354, "grad_norm": 0.057132475078105927, "learning_rate": 7.04e-07, "loss": -0.0, "num_tokens": 1968241.0, "reward": 0.2757822275161743, "reward_std": 0.4282113015651703, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11406249552965164, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8812499642372131, "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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 450.75, "completions/mean_terminated_length": 450.75, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.0356, "grad_norm": 0.0760892853140831, "learning_rate": 7.079999999999999e-07, "loss": 0.0, "num_tokens": 1979517.0, "reward": 0.5689760446548462, "reward_std": 0.6792660355567932, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.32545000314712524, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9762499928474426, "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": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 433.375, "completions/mean_terminated_length": 433.375, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.0358, "grad_norm": 0.08580948412418365, "learning_rate": 7.119999999999999e-07, "loss": -0.0, "num_tokens": 1991155.0, "reward": 0.637800931930542, "reward_std": 0.6421893835067749, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33810001611709595, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9362499713897705, "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": 0.0, "completions/max_length": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 279.1875, "completions/mean_terminated_length": 279.1875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.036, "grad_norm": 0.13242332637310028, "learning_rate": 7.159999999999999e-07, "loss": -0.0, "num_tokens": 1999974.0, "reward": 1.0306262969970703, "reward_std": 0.7176349759101868, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6862499713897705, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 357.1875, "completions/mean_terminated_length": 357.1875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.0362, "grad_norm": 0.06572093069553375, "learning_rate": 7.2e-07, "loss": -0.0, "num_tokens": 2009849.0, "reward": 0.35890740156173706, "reward_std": 0.40339112281799316, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09281250089406967, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9187500476837158, "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": 976.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 614.4375, "completions/mean_terminated_length": 614.4375, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.0364, "grad_norm": 0.036240506917238235, "learning_rate": 7.24e-07, "loss": 0.0, "num_tokens": 2023568.0, "reward": 0.2637508809566498, "reward_std": 0.36345332860946655, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1525000035762787, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.699999988079071, "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": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 291.9375, "completions/mean_terminated_length": 291.9375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.0366, "grad_norm": 0.12820208072662354, "learning_rate": 7.28e-07, "loss": 0.0, "num_tokens": 2032559.0, "reward": 0.8745981454849243, "reward_std": 0.7412576675415039, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5616937279701233, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.971875011920929, "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": 508.0, "completions/max_terminated_length": 508.0, "completions/mean_length": 326.0, "completions/mean_terminated_length": 326.0, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0368, "grad_norm": 0.09189960360527039, "learning_rate": 7.319999999999999e-07, "loss": -0.0, "num_tokens": 2041487.0, "reward": 0.3319416046142578, "reward_std": 0.4131383001804352, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10138124227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 264.5625, "completions/mean_terminated_length": 264.5625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.037, "grad_norm": 0.14386481046676636, "learning_rate": 7.359999999999999e-07, "loss": 0.0, "num_tokens": 2049480.0, "reward": 0.6560074090957642, "reward_std": 0.7682359218597412, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4370124936103821, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9793750047683716, "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": 0.0, "completions/max_length": 240.0, "completions/max_terminated_length": 240.0, "completions/mean_length": 193.8125, "completions/mean_terminated_length": 193.8125, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.0372, "grad_norm": 0.08803083747625351, "learning_rate": 7.4e-07, "loss": -0.0, "num_tokens": 2056277.0, "reward": 1.3439700603485107, "reward_std": 0.4351760745048523, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8754374980926514, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.981249988079071, "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": 0.0, "completions/max_length": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 601.6875, "completions/mean_terminated_length": 601.6875, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.0374, "grad_norm": 0.05986401438713074, "learning_rate": 7.44e-07, "loss": 0.0, "num_tokens": 2070720.0, "reward": 0.2680477499961853, "reward_std": 0.42072930932044983, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09859375655651093, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.862500011920929, "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": 756.0, "completions/max_terminated_length": 756.0, "completions/mean_length": 366.0625, "completions/mean_terminated_length": 366.0625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0376, "grad_norm": 0.12068089842796326, "learning_rate": 7.48e-07, "loss": -0.0, "num_tokens": 2080209.0, "reward": 0.5028918981552124, "reward_std": 0.5693369507789612, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25578123331069946, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8687499761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 590.0625, "completions/mean_terminated_length": 561.1333618164062, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.0378, "grad_norm": 0.045261867344379425, "learning_rate": 7.52e-07, "loss": 0.0, "num_tokens": 2094482.0, "reward": 0.3021104633808136, "reward_std": 0.4190751612186432, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1042187437415123, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8343750238418579, "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": 0.0, "completions/max_length": 312.0, "completions/max_terminated_length": 312.0, "completions/mean_length": 235.4375, "completions/mean_terminated_length": 235.4375, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.038, "grad_norm": 0.13196101784706116, "learning_rate": 7.559999999999999e-07, "loss": 0.0, "num_tokens": 2102009.0, "reward": 1.0629420280456543, "reward_std": 0.6794048547744751, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6883811950683594, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9862499833106995, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 566.1875, "completions/mean_terminated_length": 535.6666870117188, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.0382, "grad_norm": 0.05577923357486725, "learning_rate": 7.599999999999999e-07, "loss": -0.0, "num_tokens": 2115052.0, "reward": 0.3179693818092346, "reward_std": 0.44169262051582336, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26093751192092896, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.3968749940395355, "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": 0.0, "completions/max_length": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 301.75, "completions/mean_terminated_length": 301.75, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0384, "grad_norm": 0.10017109662294388, "learning_rate": 7.64e-07, "loss": 0.0, "num_tokens": 2123480.0, "reward": 0.47773247957229614, "reward_std": 0.5623816847801208, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2054625004529953, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 292.0, "completions/max_terminated_length": 292.0, "completions/mean_length": 239.8125, "completions/mean_terminated_length": 239.8125, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0386, "grad_norm": 0.17794230580329895, "learning_rate": 7.68e-07, "loss": 0.0, "num_tokens": 2131173.0, "reward": 0.6902358531951904, "reward_std": 0.7489261031150818, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4429687261581421, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9775000214576721, "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": 755.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 512.625, "completions/mean_terminated_length": 512.625, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.0388, "grad_norm": 0.06626477837562561, "learning_rate": 7.72e-07, "loss": 0.0, "num_tokens": 2143743.0, "reward": 0.4483480453491211, "reward_std": 0.5127907991409302, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20919373631477356, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9175000190734863, "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": 438.0, "completions/max_terminated_length": 438.0, "completions/mean_length": 294.3125, "completions/mean_terminated_length": 294.3125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.039, "grad_norm": 0.060651469975709915, "learning_rate": 7.76e-07, "loss": -0.0, "num_tokens": 2153012.0, "reward": 0.4643480181694031, "reward_std": 0.36508962512016296, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1161937490105629, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9675000309944153, "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": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 541.9375, "completions/mean_terminated_length": 541.9375, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0392, "grad_norm": 0.08764442056417465, "learning_rate": 7.799999999999999e-07, "loss": 0.0, "num_tokens": 2165811.0, "reward": 0.40898552536964417, "reward_std": 0.6638191938400269, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2554687559604645, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 0.0, "completions/max_length": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 360.625, "completions/mean_terminated_length": 360.625, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.0394, "grad_norm": 0.04764646291732788, "learning_rate": 7.84e-07, "loss": 0.0, "num_tokens": 2175517.0, "reward": 0.20953285694122314, "reward_std": 0.2851693034172058, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04406249523162842, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.90625, "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": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 527.625, "completions/mean_terminated_length": 527.625, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0396, "grad_norm": 0.053040698170661926, "learning_rate": 7.88e-07, "loss": -0.0, "num_tokens": 2188071.0, "reward": 0.34873834252357483, "reward_std": 0.38142821192741394, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19747498631477356, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7268750071525574, "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": 805.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 428.0, "completions/mean_terminated_length": 428.0, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.0398, "grad_norm": 0.05956544354557991, "learning_rate": 7.92e-07, "loss": 0.0, "num_tokens": 2200295.0, "reward": 0.32429826259613037, "reward_std": 0.5177905559539795, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1485937535762787, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.815625011920929, "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": 693.0, "completions/max_terminated_length": 693.0, "completions/mean_length": 488.1875, "completions/mean_terminated_length": 488.1875, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.04, "grad_norm": 0.05395005643367767, "learning_rate": 7.96e-07, "loss": 0.0, "num_tokens": 2212026.0, "reward": 0.22992271184921265, "reward_std": 0.3688678443431854, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14734375476837158, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.574999988079071, "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": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 300.5, "completions/mean_terminated_length": 300.5, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.0402, "grad_norm": 0.08739028126001358, "learning_rate": 8e-07, "loss": 0.0, "num_tokens": 2220690.0, "reward": 0.47296079993247986, "reward_std": 0.5620365738868713, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1959187388420105, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.984375, "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": 643.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 443.875, "completions/mean_terminated_length": 443.875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.0404, "grad_norm": 0.07621358335018158, "learning_rate": 8.04e-07, "loss": -0.0, "num_tokens": 2231776.0, "reward": 0.47802287340164185, "reward_std": 0.5299331545829773, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2685437500476837, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7481250166893005, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 663.8125, "completions/mean_terminated_length": 639.800048828125, "completions/min_length": 442.0, "completions/min_terminated_length": 442.0, "epoch": 0.0406, "grad_norm": 0.03625255078077316, "learning_rate": 8.08e-07, "loss": 0.0, "num_tokens": 2246317.0, "reward": 0.2712480127811432, "reward_std": 0.42374369502067566, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10499374568462372, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8806250095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 529.8125, "completions/mean_terminated_length": 496.86669921875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0408, "grad_norm": 0.08742740750312805, "learning_rate": 8.12e-07, "loss": -0.0, "num_tokens": 2259194.0, "reward": 0.4682043790817261, "reward_std": 0.6437572836875916, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.24890625476837158, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8468749523162842, "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": 904.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 464.875, "completions/mean_terminated_length": 464.875, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.041, "grad_norm": 0.11402137577533722, "learning_rate": 8.159999999999999e-07, "loss": -0.0, "num_tokens": 2270184.0, "reward": 0.6650010347366333, "reward_std": 0.6980466842651367, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39249998331069946, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 530.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 386.75, "completions/mean_terminated_length": 386.75, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.0412, "grad_norm": 0.05152532830834389, "learning_rate": 8.199999999999999e-07, "loss": 0.0, "num_tokens": 2279972.0, "reward": 0.39214783906936646, "reward_std": 0.3935956060886383, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09679350256919861, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9629937410354614, "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": 0.0, "completions/max_length": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 400.4375, "completions/mean_terminated_length": 400.4375, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0414, "grad_norm": 0.03376661241054535, "learning_rate": 8.24e-07, "loss": -0.0, "num_tokens": 2290667.0, "reward": 0.1305229663848877, "reward_std": 0.23365728557109833, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.011043749749660492, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.9100000262260437, "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": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 484.6875, "completions/mean_terminated_length": 484.6875, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.0416, "grad_norm": 0.07746847718954086, "learning_rate": 8.28e-07, "loss": 0.0, "num_tokens": 2302070.0, "reward": 0.4091354310512543, "reward_std": 0.6640820503234863, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.255768746137619, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 365.0, "completions/max_terminated_length": 365.0, "completions/mean_length": 297.6875, "completions/mean_terminated_length": 297.6875, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.0418, "grad_norm": 0.08926597982645035, "learning_rate": 8.319999999999999e-07, "loss": -0.0, "num_tokens": 2310625.0, "reward": 0.4891383945941925, "reward_std": 0.4948078393936157, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22827500104904175, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8450000286102295, "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": 0.0, "completions/max_length": 592.0, "completions/max_terminated_length": 592.0, "completions/mean_length": 427.6875, "completions/mean_terminated_length": 427.6875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.042, "grad_norm": 0.09087108820676804, "learning_rate": 8.359999999999999e-07, "loss": 0.0, "num_tokens": 2321308.0, "reward": 0.5652887225151062, "reward_std": 0.6795036792755127, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3180750012397766, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8443750143051147, "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": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 522.875, "completions/mean_terminated_length": 522.875, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.0422, "grad_norm": 0.0436578132212162, "learning_rate": 8.399999999999999e-07, "loss": 0.0, "num_tokens": 2333434.0, "reward": 0.259063720703125, "reward_std": 0.41439250111579895, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08062499761581421, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.903124988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 452.6875, "completions/mean_terminated_length": 414.60003662109375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.0424, "grad_norm": 0.09693451970815659, "learning_rate": 8.439999999999999e-07, "loss": 0.0, "num_tokens": 2344405.0, "reward": 0.7660039663314819, "reward_std": 0.6253086924552917, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40700626373291016, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8331249952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 543.0, "completions/mean_terminated_length": 510.933349609375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.0426, "grad_norm": 0.05475315824151039, "learning_rate": 8.48e-07, "loss": 0.0, "num_tokens": 2357541.0, "reward": 0.320313423871994, "reward_std": 0.4422723054885864, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.265625, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.4906250238418579, "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": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 496.3125, "completions/mean_terminated_length": 496.3125, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.0428, "grad_norm": 0.03649654984474182, "learning_rate": 8.52e-07, "loss": 0.0, "num_tokens": 2369962.0, "reward": 0.20273546874523163, "reward_std": 0.27055659890174866, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.71875, "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": 714.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 537.5625, "completions/mean_terminated_length": 537.5625, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.043, "grad_norm": 0.07135090976953506, "learning_rate": 8.559999999999999e-07, "loss": -0.0, "num_tokens": 2382083.0, "reward": 0.28109508752822876, "reward_std": 0.6043334603309631, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.18718749284744263, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9087499976158142, "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": 368.0, "completions/max_terminated_length": 368.0, "completions/mean_length": 287.25, "completions/mean_terminated_length": 287.25, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.0432, "grad_norm": 0.07327918708324432, "learning_rate": 8.599999999999999e-07, "loss": 0.0, "num_tokens": 2390599.0, "reward": 0.37959787249565125, "reward_std": 0.41544950008392334, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13419374823570251, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9300000071525574, "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": 0.0, "completions/max_length": 362.0, "completions/max_terminated_length": 362.0, "completions/mean_length": 267.6875, "completions/mean_terminated_length": 267.6875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.0434, "grad_norm": 0.14973832666873932, "learning_rate": 8.639999999999999e-07, "loss": -0.0, "num_tokens": 2398898.0, "reward": 1.1216166019439697, "reward_std": 0.6688916683197021, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7432312369346619, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9462500214576721, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 511.25, "completions/mean_terminated_length": 511.25, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.0436, "grad_norm": 0.025454053655266762, "learning_rate": 8.68e-07, "loss": -0.0, "num_tokens": 2411286.0, "reward": 0.1599203646183014, "reward_std": 0.24521981179714203, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.007337499875575304, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9856250286102295, "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": 0.0, "completions/max_length": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 482.5, "completions/mean_terminated_length": 482.5, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.0438, "grad_norm": 0.054896045476198196, "learning_rate": 8.72e-07, "loss": 0.0, "num_tokens": 2423038.0, "reward": 0.39138221740722656, "reward_std": 0.5046852231025696, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15776249766349792, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9031250476837158, "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": 0.0, "completions/max_length": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 528.8125, "completions/mean_terminated_length": 528.8125, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.044, "grad_norm": 0.06159689277410507, "learning_rate": 8.76e-07, "loss": -0.0, "num_tokens": 2436491.0, "reward": 0.41804802417755127, "reward_std": 0.5253297090530396, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2110937535762787, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.731249988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 510.4375, "completions/mean_terminated_length": 476.20001220703125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.0442, "grad_norm": 0.056172896176576614, "learning_rate": 8.799999999999999e-07, "loss": -0.0, "num_tokens": 2449522.0, "reward": 0.3610037863254547, "reward_std": 0.3577970266342163, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15950624644756317, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5049999952316284, "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": 0.0, "completions/max_length": 417.0, "completions/max_terminated_length": 417.0, "completions/mean_length": 285.375, "completions/mean_terminated_length": 285.375, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.0444, "grad_norm": 0.05109907686710358, "learning_rate": 8.839999999999999e-07, "loss": -0.0, "num_tokens": 2458056.0, "reward": 0.37711039185523987, "reward_std": 0.26315781474113464, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06671874970197678, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8949999809265137, "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": 324.0, "completions/max_terminated_length": 324.0, "completions/mean_length": 206.0625, "completions/mean_terminated_length": 206.0625, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.0446, "grad_norm": 0.14047451317310333, "learning_rate": 8.88e-07, "loss": -0.0, "num_tokens": 2464729.0, "reward": 0.791532576084137, "reward_std": 0.599698543548584, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3955624997615814, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9825000166893005, "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": 0.0, "completions/max_length": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 508.75, "completions/mean_terminated_length": 508.75, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.0448, "grad_norm": 0.05510563775897026, "learning_rate": 8.92e-07, "loss": 0.0, "num_tokens": 2476725.0, "reward": 0.4111577272415161, "reward_std": 0.5846126675605774, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.19731250405311584, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9862500429153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 639.0, "completions/mean_terminated_length": 613.3333740234375, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.045, "grad_norm": 0.04691677540540695, "learning_rate": 8.96e-07, "loss": 0.0, "num_tokens": 2490821.0, "reward": 0.19945421814918518, "reward_std": 0.4136039912700653, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08640625327825546, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8987500071525574, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 630.25, "completions/mean_terminated_length": 604.0000610351562, "completions/min_length": 458.0, "completions/min_terminated_length": 458.0, "epoch": 0.0452, "grad_norm": 0.03584437072277069, "learning_rate": 9e-07, "loss": -0.0, "num_tokens": 2505577.0, "reward": 0.17937608063220978, "reward_std": 0.3406701982021332, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10875000059604645, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.7631250023841858, "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": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 312.125, "completions/mean_terminated_length": 312.125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.0454, "grad_norm": 0.13080304861068726, "learning_rate": 9.039999999999999e-07, "loss": 0.0, "num_tokens": 2514715.0, "reward": 0.8776576519012451, "reward_std": 0.7397249341011047, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5678125023841858, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9737499952316284, "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": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 325.5625, "completions/mean_terminated_length": 325.5625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.0456, "grad_norm": 0.1515752673149109, "learning_rate": 9.08e-07, "loss": -0.0, "num_tokens": 2524132.0, "reward": 0.5963946580886841, "reward_std": 0.7344302535057068, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3802874982357025, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 303.0, "completions/max_terminated_length": 303.0, "completions/mean_length": 258.125, "completions/mean_terminated_length": 258.125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0458, "grad_norm": 0.1228063777089119, "learning_rate": 9.12e-07, "loss": -0.0, "num_tokens": 2532070.0, "reward": 0.7534981369972229, "reward_std": 0.7056349515914917, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.44449374079704285, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 451.0, "completions/max_terminated_length": 451.0, "completions/mean_length": 263.0625, "completions/mean_terminated_length": 263.0625, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.046, "grad_norm": 0.10843759775161743, "learning_rate": 9.16e-07, "loss": -0.0, "num_tokens": 2540039.0, "reward": 0.3731575012207031, "reward_std": 0.41456031799316406, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12131249904632568, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9493750333786011, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 490.125, "completions/mean_terminated_length": 413.8571472167969, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.0462, "grad_norm": 0.034995391964912415, "learning_rate": 9.2e-07, "loss": 0.0, "num_tokens": 2551433.0, "reward": 0.16539166867733002, "reward_std": 0.2537343204021454, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8456250429153442, "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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 261.0, "completions/mean_terminated_length": 261.0, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.0464, "grad_norm": 0.13089807331562042, "learning_rate": 9.24e-07, "loss": 0.0, "num_tokens": 2559369.0, "reward": 0.5109982490539551, "reward_std": 0.5480625033378601, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2094937413930893, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9212499856948853, "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": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 457.4375, "completions/mean_terminated_length": 457.4375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.0466, "grad_norm": 0.04172377660870552, "learning_rate": 9.28e-07, "loss": 0.0, "num_tokens": 2570784.0, "reward": 0.35132598876953125, "reward_std": 0.2842935621738434, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07764999568462372, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8462499380111694, "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": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 615.375, "completions/mean_terminated_length": 615.375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.0468, "grad_norm": 0.04014880210161209, "learning_rate": 9.32e-07, "loss": 0.0, "num_tokens": 2584630.0, "reward": 0.16218885779380798, "reward_std": 0.3091006875038147, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07437500357627869, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8456249833106995, "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": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 326.25, "completions/mean_terminated_length": 326.25, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.047, "grad_norm": 0.08564586937427521, "learning_rate": 9.36e-07, "loss": 0.0, "num_tokens": 2594250.0, "reward": 0.4751541018486023, "reward_std": 0.6447211503982544, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.262806236743927, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 0.0, "completions/max_length": 476.0, "completions/max_terminated_length": 476.0, "completions/mean_length": 330.4375, "completions/mean_terminated_length": 330.4375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.0472, "grad_norm": 0.12743502855300903, "learning_rate": 9.399999999999999e-07, "loss": -0.0, "num_tokens": 2603985.0, "reward": 0.5624233484268188, "reward_std": 0.7498961091041565, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3748437464237213, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9806250333786011, "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": 388.0, "completions/max_terminated_length": 388.0, "completions/mean_length": 297.9375, "completions/mean_terminated_length": 297.9375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0474, "grad_norm": 0.12305436283349991, "learning_rate": 9.439999999999999e-07, "loss": 0.0, "num_tokens": 2612400.0, "reward": 0.9716886878013611, "reward_std": 0.7159633636474609, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6308750510215759, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 0.0, "completions/max_length": 768.0, "completions/max_terminated_length": 768.0, "completions/mean_length": 619.8125, "completions/mean_terminated_length": 619.8125, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.0476, "grad_norm": 0.043658025562763214, "learning_rate": 9.479999999999999e-07, "loss": 0.0, "num_tokens": 2626573.0, "reward": 0.17296646535396576, "reward_std": 0.41544389724731445, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.09593124687671661, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.6493749618530273, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 607.0, "completions/mean_length": 498.1875, "completions/mean_terminated_length": 423.0714416503906, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.0478, "grad_norm": 0.08782634139060974, "learning_rate": 9.52e-07, "loss": 0.0, "num_tokens": 2638384.0, "reward": 0.3116418123245239, "reward_std": 0.600275456905365, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.18578125536441803, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.6593749523162842, "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": 323.0, "completions/max_terminated_length": 323.0, "completions/mean_length": 268.3125, "completions/mean_terminated_length": 268.3125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.048, "grad_norm": 0.1418672800064087, "learning_rate": 9.559999999999998e-07, "loss": 0.0, "num_tokens": 2646517.0, "reward": 0.6627297401428223, "reward_std": 0.6991822123527527, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3879562318325043, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9268749952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 474.3125, "completions/mean_terminated_length": 437.66668701171875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.0482, "grad_norm": 0.051750510931015015, "learning_rate": 9.6e-07, "loss": -0.0, "num_tokens": 2658426.0, "reward": 0.29648566246032715, "reward_std": 0.2707010805606842, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8500000238418579, "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": 178.0, "completions/max_terminated_length": 178.0, "completions/mean_length": 127.0625, "completions/mean_terminated_length": 127.0625, "completions/min_length": 91.0, "completions/min_terminated_length": 91.0, "epoch": 0.0484, "grad_norm": 0.2754916846752167, "learning_rate": 9.64e-07, "loss": -0.0, "num_tokens": 2664235.0, "reward": 0.5937486886978149, "reward_std": 0.663791298866272, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3124937415122986, "rewards/confidence_one_or_zero": 0.875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9987499713897705, "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": 244.0, "completions/max_terminated_length": 244.0, "completions/mean_length": 175.6875, "completions/mean_terminated_length": 175.6875, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 0.0486, "grad_norm": 0.16145707666873932, "learning_rate": 9.679999999999999e-07, "loss": -0.0, "num_tokens": 2670870.0, "reward": 0.6953763961791992, "reward_std": 0.5995557904243469, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.328249990940094, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9200000166893005, "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": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 350.75, "completions/mean_terminated_length": 350.75, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.0488, "grad_norm": 0.14809547364711761, "learning_rate": 9.72e-07, "loss": -0.0, "num_tokens": 2680466.0, "reward": 0.690157413482666, "reward_std": 0.7488361597061157, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4428125023841858, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 959.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 693.5625, "completions/mean_terminated_length": 693.5625, "completions/min_length": 523.0, "completions/min_terminated_length": 523.0, "epoch": 0.049, "grad_norm": 0.039376288652420044, "learning_rate": 9.759999999999998e-07, "loss": 0.0, "num_tokens": 2695371.0, "reward": 0.2217199057340622, "reward_std": 0.5170965194702148, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13093750178813934, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 635.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 453.3125, "completions/mean_terminated_length": 453.3125, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.0492, "grad_norm": 0.0847303569316864, "learning_rate": 9.8e-07, "loss": -0.0, "num_tokens": 2706736.0, "reward": 0.4660949110984802, "reward_std": 0.5328766107559204, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1821874976158142, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9200000166893005, "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": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 423.875, "completions/mean_terminated_length": 423.875, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.0494, "grad_norm": 0.051883265376091, "learning_rate": 9.84e-07, "loss": 0.0, "num_tokens": 2717198.0, "reward": 0.47997623682022095, "reward_std": 0.3867835998535156, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14745000004768372, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9156249761581421, "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": 0.0, "completions/max_length": 705.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 500.75, "completions/mean_terminated_length": 500.75, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.0496, "grad_norm": 0.05023181438446045, "learning_rate": 9.88e-07, "loss": -0.0, "num_tokens": 2729402.0, "reward": 0.29343897104263306, "reward_std": 0.41447290778160095, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08687499165534973, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8581249713897705, "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": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 485.25, "completions/mean_terminated_length": 485.25, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.0498, "grad_norm": 0.04993528500199318, "learning_rate": 9.92e-07, "loss": -0.0, "num_tokens": 2742382.0, "reward": 0.42997613549232483, "reward_std": 0.3817908465862274, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1099499985575676, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 847.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 550.6875, "completions/mean_terminated_length": 550.6875, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.05, "grad_norm": 0.06858085095882416, "learning_rate": 9.959999999999999e-07, "loss": -0.0, "num_tokens": 2755097.0, "reward": 0.35242271423339844, "reward_std": 0.5415217876434326, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20484375953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8012499809265137, "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": 301.0, "completions/max_terminated_length": 301.0, "completions/mean_length": 250.1875, "completions/mean_terminated_length": 250.1875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.0502, "grad_norm": 0.09872546792030334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2762732.0, "reward": 0.4442419111728668, "reward_std": 0.4797784984111786, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1384812444448471, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9825000166893005, "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": 921.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 418.875, "completions/mean_terminated_length": 418.875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.0504, "grad_norm": 0.08761284500360489, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2773866.0, "reward": 0.7003666162490845, "reward_std": 0.6762117743492126, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40073126554489136, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8912500143051147, "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": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 356.0, "completions/mean_terminated_length": 356.0, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.0506, "grad_norm": 0.11257665604352951, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2784346.0, "reward": 0.5534260869026184, "reward_std": 0.5302914977073669, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23184999823570251, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 0.0, "completions/max_length": 200.0, "completions/max_terminated_length": 200.0, "completions/mean_length": 144.9375, "completions/mean_terminated_length": 144.9375, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "epoch": 0.0508, "grad_norm": 0.119099460542202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2790537.0, "reward": 1.3435702323913574, "reward_std": 0.4365370273590088, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8746374845504761, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9900000095367432, "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": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 430.4375, "completions/mean_terminated_length": 430.4375, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.051, "grad_norm": 0.14179790019989014, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2801264.0, "reward": 0.6872514486312866, "reward_std": 0.7497110366821289, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.43699997663497925, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8575000166893005, "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": 275.0, "completions/max_terminated_length": 275.0, "completions/mean_length": 168.75, "completions/mean_terminated_length": 168.75, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "epoch": 0.0512, "grad_norm": 0.12870573997497559, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2807580.0, "reward": 0.9593759179115295, "reward_std": 0.46741437911987305, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.731249988079071, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.38749998807907104, "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": 791.0, "completions/max_terminated_length": 791.0, "completions/mean_length": 478.8125, "completions/mean_terminated_length": 478.8125, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.0514, "grad_norm": 0.03609219193458557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2819081.0, "reward": 0.34296977519989014, "reward_std": 0.27437490224838257, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06093749776482582, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.953125, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 688.0625, "completions/mean_terminated_length": 640.0714721679688, "completions/min_length": 534.0, "completions/min_terminated_length": 534.0, "epoch": 0.0516, "grad_norm": 0.06985849142074585, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2834154.0, "reward": 0.5168758630752563, "reward_std": 0.6596401333808899, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3462499976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.734375, "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": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 336.125, "completions/mean_terminated_length": 336.125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.0518, "grad_norm": 0.044154778122901917, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2843180.0, "reward": 0.4115636348724365, "reward_std": 0.2454081028699875, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07312499731779099, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9593749642372131, "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": 0.0, "completions/max_length": 742.0, "completions/max_terminated_length": 742.0, "completions/mean_length": 554.25, "completions/mean_terminated_length": 554.25, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.052, "grad_norm": 0.07736440747976303, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2856176.0, "reward": 0.4684359133243561, "reward_std": 0.7175864577293396, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3118687570095062, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9806250333786011, "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": 0.0, "completions/max_length": 964.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 712.25, "completions/mean_terminated_length": 712.25, "completions/min_length": 547.0, "completions/min_terminated_length": 547.0, "epoch": 0.0522, "grad_norm": 0.0539529025554657, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2871508.0, "reward": 0.18734465539455414, "reward_std": 0.5119205713272095, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12468750029802322, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6612499952316284, "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": 300.0, "completions/max_terminated_length": 300.0, "completions/mean_length": 244.25, "completions/mean_terminated_length": 244.25, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.0524, "grad_norm": 0.1577993482351303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2879368.0, "reward": 0.7497607469558716, "reward_std": 0.7068345546722412, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4370187222957611, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.981249988079071, "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": 994.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 804.875, "completions/mean_terminated_length": 804.875, "completions/min_length": 599.0, "completions/min_terminated_length": 599.0, "epoch": 0.0526, "grad_norm": 0.02301195077598095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2896758.0, "reward": 0.13414160907268524, "reward_std": 0.24020564556121826, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.859375, "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": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 325.8125, "completions/mean_terminated_length": 325.8125, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.0528, "grad_norm": 0.08761764317750931, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2905523.0, "reward": 1.074051022529602, "reward_std": 0.5897222757339478, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6481000185012817, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 623.0, "completions/mean_terminated_length": 489.3333435058594, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.053, "grad_norm": 0.03916896507143974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2919139.0, "reward": 0.26156359910964966, "reward_std": 0.3269492983818054, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08562500774860382, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.5625, "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": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 413.25, "completions/mean_terminated_length": 413.25, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.0532, "grad_norm": 0.07530451565980911, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2930055.0, "reward": 0.2278793305158615, "reward_std": 0.4140791893005371, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08075624704360962, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9725000262260437, "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": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 613.1875, "completions/mean_terminated_length": 613.1875, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.0534, "grad_norm": 0.05499095842242241, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2943722.0, "reward": 0.18711033463478088, "reward_std": 0.5112805962562561, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12421874701976776, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.8343749642372131, "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": 0.0, "completions/max_length": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 199.8125, "completions/mean_terminated_length": 199.8125, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "epoch": 0.0536, "grad_norm": 0.11870162189006805, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2950439.0, "reward": 1.3152886629104614, "reward_std": 0.39656686782836914, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8180750012397766, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9825000166893005, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 740.25, "completions/mean_terminated_length": 645.6666870117188, "completions/min_length": 449.0, "completions/min_terminated_length": 449.0, "epoch": 0.0538, "grad_norm": 0.023032590746879578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2966363.0, "reward": 0.13356958329677582, "reward_std": 0.23905199766159058, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.017137499526143074, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.5256249904632568, "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": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 400.375, "completions/mean_terminated_length": 400.375, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.054, "grad_norm": 0.08284304291009903, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 2977121.0, "reward": 0.3539729118347168, "reward_std": 0.5106425881385803, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14544373750686646, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.8387500047683716, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 838.625, "completions/mean_terminated_length": 795.84619140625, "completions/min_length": 475.0, "completions/min_terminated_length": 475.0, "epoch": 0.0542, "grad_norm": 0.05170236900448799, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 2995803.0, "reward": 0.43085983395576477, "reward_std": 0.534288763999939, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3617187440395355, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.234375, "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": 0.0, "completions/max_length": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 271.75, "completions/mean_terminated_length": 271.75, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.0544, "grad_norm": 0.10331044346094131, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3004583.0, "reward": 1.1909699440002441, "reward_std": 0.5690386295318604, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7569375038146973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9825000166893005, "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": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 497.875, "completions/mean_terminated_length": 497.875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.0546, "grad_norm": 0.03607700392603874, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3016101.0, "reward": 0.2027355283498764, "reward_std": 0.27055656909942627, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.9612499475479126, "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": 414.0, "completions/max_terminated_length": 414.0, "completions/mean_length": 306.125, "completions/mean_terminated_length": 306.125, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.0548, "grad_norm": 0.12310373038053513, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3024631.0, "reward": 0.630776047706604, "reward_std": 0.7174456715583801, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38655000925064087, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9037500023841858, "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": 0.0, "completions/max_length": 336.0, "completions/max_terminated_length": 336.0, "completions/mean_length": 273.75, "completions/mean_terminated_length": 273.75, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.055, "grad_norm": 0.10095721483230591, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3032915.0, "reward": 0.8279260396957397, "reward_std": 0.5642293691635132, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40584999322891235, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 388.0, "completions/max_terminated_length": 388.0, "completions/mean_length": 324.5, "completions/mean_terminated_length": 324.5, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.0552, "grad_norm": 0.0959453210234642, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3042635.0, "reward": 0.6631885170936584, "reward_std": 0.6984014511108398, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38887500762939453, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9618750214576721, "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": 0.0, "completions/max_length": 752.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 359.5625, "completions/mean_terminated_length": 359.5625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0554, "grad_norm": 0.06353491544723511, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3053044.0, "reward": 0.4942728281021118, "reward_std": 0.4666177034378052, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17604374885559082, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.956250011920929, "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": 0.0, "completions/max_length": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 501.3125, "completions/mean_terminated_length": 501.3125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0556, "grad_norm": 0.07936285436153412, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3064617.0, "reward": 0.5475791096687317, "reward_std": 0.6952944397926331, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3451562523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9443749785423279, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 561.125, "completions/mean_terminated_length": 530.2667236328125, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.0558, "grad_norm": 0.06080600246787071, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3077211.0, "reward": 0.34988877177238464, "reward_std": 0.5093868970870972, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13727501034736633, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9225000143051147, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 697.5625, "completions/mean_terminated_length": 675.800048828125, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.056, "grad_norm": 0.045243117958307266, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3092388.0, "reward": 0.28257906436920166, "reward_std": 0.4390445053577423, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1276562511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7593749761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 692.625, "completions/mean_terminated_length": 670.5333862304688, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.0562, "grad_norm": 0.07109174132347107, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3107454.0, "reward": 0.4684387147426605, "reward_std": 0.7175915837287903, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.31187498569488525, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.9149999618530273, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 294.8125, "completions/mean_terminated_length": 294.8125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.0564, "grad_norm": 0.1321430802345276, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3115899.0, "reward": 0.7524980306625366, "reward_std": 0.6816515326499939, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.504993736743927, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7362500429153442, "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": 588.0, "completions/mean_length": 364.375, "completions/mean_terminated_length": 320.4000244140625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0566, "grad_norm": 0.09657549858093262, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3125633.0, "reward": 0.7006980180740356, "reward_std": 0.5976478457450867, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33889374136924744, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9081249833106995, "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": 0.0, "completions/max_length": 391.0, "completions/max_terminated_length": 391.0, "completions/mean_length": 330.75, "completions/mean_terminated_length": 330.75, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.0568, "grad_norm": 0.13704033195972443, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3134701.0, "reward": 0.6367980241775513, "reward_std": 0.6427855491638184, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3360937833786011, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 457.125, "completions/mean_terminated_length": 457.125, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.057, "grad_norm": 0.060158610343933105, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3145711.0, "reward": 0.4930354654788971, "reward_std": 0.4665716290473938, "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": 0.9606249928474426, "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": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 526.375, "completions/mean_terminated_length": 526.375, "completions/min_length": 379.0, "completions/min_terminated_length": 379.0, "epoch": 0.0572, "grad_norm": 0.031361598521471024, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3157845.0, "reward": 0.38015708327293396, "reward_std": 0.26495176553726196, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07281249761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8374999761581421, "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": 0.0, "completions/max_length": 1010.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 574.375, "completions/mean_terminated_length": 574.375, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.0574, "grad_norm": 0.055613864213228226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3171099.0, "reward": 0.5848442912101746, "reward_std": 0.43788689374923706, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.48218750953674316, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.4781250059604645, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 555.375, "completions/mean_terminated_length": 524.1333618164062, "completions/min_length": 374.0, "completions/min_terminated_length": 374.0, "epoch": 0.0576, "grad_norm": 0.03530701622366905, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3183601.0, "reward": 0.4139852523803711, "reward_std": 0.3082055449485779, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1404687464237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7749999761581421, "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": 383.75, "completions/mean_terminated_length": 383.75, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.0578, "grad_norm": 0.09499353170394897, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3193821.0, "reward": 0.7258418798446655, "reward_std": 0.6547768712043762, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3891812562942505, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9750000238418579, "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": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 293.75, "completions/mean_terminated_length": 293.75, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.058, "grad_norm": 0.11191976815462112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3202713.0, "reward": 1.196403980255127, "reward_std": 0.5560608506202698, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7678062319755554, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8462499976158142, "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": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 570.25, "completions/mean_terminated_length": 570.25, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.0582, "grad_norm": 0.04049906134605408, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3215789.0, "reward": 0.12796969711780548, "reward_std": 0.3903299570083618, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.06843750178813934, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9431250095367432, "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": 863.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 351.6875, "completions/mean_terminated_length": 351.6875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.0584, "grad_norm": 0.06625690311193466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3225592.0, "reward": 0.36984479427337646, "reward_std": 0.3504473865032196, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1771874874830246, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.78125, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 473.5625, "completions/mean_terminated_length": 394.9285888671875, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.0586, "grad_norm": 0.05301019921898842, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3237201.0, "reward": 0.36765697598457336, "reward_std": 0.36405786871910095, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17281249165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.640625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 627.375, "completions/mean_terminated_length": 600.933349609375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.0588, "grad_norm": 0.06339131295681, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3251319.0, "reward": 0.36851662397384644, "reward_std": 0.519282877445221, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1745312511920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.762499988079071, "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": 851.0, "completions/max_terminated_length": 851.0, "completions/mean_length": 582.25, "completions/mean_terminated_length": 582.25, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.059, "grad_norm": 0.056916795670986176, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3264395.0, "reward": 0.12187579274177551, "reward_std": 0.33310574293136597, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11875000596046448, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.7356250286102295, "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": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 501.9375, "completions/mean_terminated_length": 501.9375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.0592, "grad_norm": 0.036013778299093246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3276298.0, "reward": 0.13414183259010315, "reward_std": 0.24020549654960632, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 357.0, "completions/max_terminated_length": 357.0, "completions/mean_length": 269.875, "completions/mean_terminated_length": 269.875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.0594, "grad_norm": 0.12173142284154892, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3284968.0, "reward": 1.124657392501831, "reward_std": 0.6706154346466064, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7493125200271606, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9812500476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 645.9375, "completions/mean_terminated_length": 620.7333374023438, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.0596, "grad_norm": 0.05635949224233627, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3299015.0, "reward": 0.32711029052734375, "reward_std": 0.5190215110778809, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15421874821186066, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8300000429153442, "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": 0.0, "completions/max_length": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 288.8125, "completions/mean_terminated_length": 288.8125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.0598, "grad_norm": 0.10022690147161484, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3307940.0, "reward": 0.48151031136512756, "reward_std": 0.5627217888832092, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2130187451839447, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9662500619888306, "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": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 371.75, "completions/mean_terminated_length": 371.75, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.06, "grad_norm": 0.133157417178154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3317728.0, "reward": 0.6288166046142578, "reward_std": 0.7175087332725525, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38263124227523804, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8956249952316284, "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": 304.0, "completions/max_terminated_length": 304.0, "completions/mean_length": 233.9375, "completions/mean_terminated_length": 233.9375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.0602, "grad_norm": 0.09722943603992462, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3325471.0, "reward": 0.9207542538642883, "reward_std": 0.5429714322090149, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4665062427520752, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9737499952316284, "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": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 378.125, "completions/mean_terminated_length": 378.125, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.0604, "grad_norm": 0.047361016273498535, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3335169.0, "reward": 0.8185161352157593, "reward_std": 0.2583368420600891, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6995312571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.40625, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 612.375, "completions/mean_terminated_length": 553.5714721679688, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.0606, "grad_norm": 0.026242706924676895, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3349223.0, "reward": 0.1028917133808136, "reward_std": 0.2212081253528595, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.018281249329447746, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.8425000309944153, "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": 884.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 410.9375, "completions/mean_terminated_length": 410.9375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.0608, "grad_norm": 0.07975900173187256, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3359638.0, "reward": 0.504048228263855, "reward_std": 0.6321870684623718, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2580937445163727, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9156249761581421, "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": 609.0, "completions/max_terminated_length": 609.0, "completions/mean_length": 436.875, "completions/mean_terminated_length": 436.875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.061, "grad_norm": 0.07088575512170792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3371348.0, "reward": 0.4305446147918701, "reward_std": 0.5304381251335144, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23608750104904175, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.8981249928474426, "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": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 522.125, "completions/mean_terminated_length": 522.125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.0612, "grad_norm": 0.06223578378558159, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3383446.0, "reward": 0.42257294058799744, "reward_std": 0.4940418303012848, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15764376521110535, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9612499475479126, "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": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 279.375, "completions/mean_terminated_length": 279.375, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0614, "grad_norm": 0.024602973833680153, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3391564.0, "reward": 1.4526417255401611, "reward_std": 0.18603846430778503, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9677811861038208, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9481250047683716, "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": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 598.25, "completions/mean_terminated_length": 598.25, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.0616, "grad_norm": 0.06400107592344284, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3405104.0, "reward": 0.31662940979003906, "reward_std": 0.5137941241264343, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.13325625658035278, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.971875011920929, "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": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 271.625, "completions/mean_terminated_length": 271.625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.0618, "grad_norm": 0.04160322993993759, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3413466.0, "reward": 0.4619823694229126, "reward_std": 0.18123719096183777, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04896249994635582, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.971875011920929, "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": 641.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 425.375, "completions/mean_terminated_length": 425.375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.062, "grad_norm": 0.0492420494556427, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3424976.0, "reward": 0.3524855077266693, "reward_std": 0.2913472354412079, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07996875047683716, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.934999942779541, "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": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 558.0625, "completions/mean_terminated_length": 558.0625, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.0622, "grad_norm": 0.030915437266230583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3438769.0, "reward": 0.34296971559524536, "reward_std": 0.27437499165534973, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06093749776482582, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.949999988079071, "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": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 355.1875, "completions/mean_terminated_length": 355.1875, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.0624, "grad_norm": 0.11046314239501953, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3448164.0, "reward": 1.0307011604309082, "reward_std": 0.7176875472068787, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6863999962806702, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 410.0, "completions/max_terminated_length": 410.0, "completions/mean_length": 329.75, "completions/mean_terminated_length": 329.75, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.0626, "grad_norm": 0.062026821076869965, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3457904.0, "reward": 0.6290446519851685, "reward_std": 0.3658413290977478, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19558748602867126, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 336.5, "completions/mean_terminated_length": 336.5, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.0628, "grad_norm": 0.08997749537229538, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3467560.0, "reward": 0.18749511241912842, "reward_std": 0.5123302936553955, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12498749792575836, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 323.875, "completions/mean_terminated_length": 323.875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.063, "grad_norm": 0.09375667572021484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3477110.0, "reward": 1.2010791301727295, "reward_std": 0.5566341280937195, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7771562337875366, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9481250047683716, "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": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 347.6875, "completions/mean_terminated_length": 347.6875, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.0632, "grad_norm": 0.13207237422466278, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3486865.0, "reward": 0.8492196798324585, "reward_std": 0.6968263983726501, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.510937511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9199999570846558, "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": 639.0, "completions/max_terminated_length": 639.0, "completions/mean_length": 412.4375, "completions/mean_terminated_length": 412.4375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.0634, "grad_norm": 0.06330056488513947, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3498264.0, "reward": 0.3274231255054474, "reward_std": 0.36913371086120605, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1548437476158142, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.793749988079071, "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": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 457.375, "completions/mean_terminated_length": 457.375, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0636, "grad_norm": 0.034283172339200974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3509838.0, "reward": 0.3771103620529175, "reward_std": 0.26315781474113464, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06671874970197678, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 348.0, "completions/max_terminated_length": 348.0, "completions/mean_length": 271.6875, "completions/mean_terminated_length": 271.6875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.0638, "grad_norm": 0.1136912927031517, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3517769.0, "reward": 1.0373165607452393, "reward_std": 0.6384738683700562, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6371312141418457, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 294.0, "completions/max_terminated_length": 294.0, "completions/mean_length": 219.8125, "completions/mean_terminated_length": 219.8125, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.064, "grad_norm": 0.11255137622356415, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3524742.0, "reward": 1.1283389329910278, "reward_std": 0.49497145414352417, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6316750049591064, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.981249988079071, "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": 863.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 585.5625, "completions/mean_terminated_length": 585.5625, "completions/min_length": 404.0, "completions/min_terminated_length": 404.0, "epoch": 0.0642, "grad_norm": 0.027323564514517784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3538879.0, "reward": 0.09984529763460159, "reward_std": 0.21490336954593658, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.012187499552965164, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.9831249713897705, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 634.5625, "completions/mean_terminated_length": 578.9285888671875, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.0644, "grad_norm": 0.03670664131641388, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3553176.0, "reward": 0.20257915556430817, "reward_std": 0.2708997428417206, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.030156249180436134, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8436875343322754, "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": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 353.9375, "completions/mean_terminated_length": 353.9375, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.0646, "grad_norm": 0.08329904824495316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3562391.0, "reward": 0.5892979502677917, "reward_std": 0.39679473638534546, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4285937547683716, "rewards/confidence_one_or_zero": 0.6875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.4468749761581421, "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": 0.0, "completions/max_length": 861.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 617.5625, "completions/mean_terminated_length": 617.5625, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.0648, "grad_norm": 0.033714212477207184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3577200.0, "reward": 0.07921954989433289, "reward_std": 0.21868064999580383, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03343749791383743, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.6743749976158142, "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": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 354.5, "completions/mean_terminated_length": 354.5, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.065, "grad_norm": 0.1271912306547165, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3587288.0, "reward": 0.8157634735107422, "reward_std": 0.7254511117935181, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5065249800682068, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 811.8125, "completions/mean_terminated_length": 741.0833740234375, "completions/min_length": 524.0, "completions/min_terminated_length": 524.0, "epoch": 0.0652, "grad_norm": 0.039745453745126724, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3605381.0, "reward": 0.32453209161758423, "reward_std": 0.5171641707420349, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14906249940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.6612499952316284, "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": 0.0, "completions/max_length": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 540.25, "completions/mean_terminated_length": 540.25, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0654, "grad_norm": 0.049348533153533936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3618409.0, "reward": 0.35265713930130005, "reward_std": 0.3293524980545044, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14281249046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8243749737739563, "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": 0.0, "completions/max_length": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 383.0, "completions/mean_terminated_length": 383.0, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.0656, "grad_norm": 0.09120383113622665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3628921.0, "reward": 0.2994542121887207, "reward_std": 0.4177921712398529, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0989062488079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 435.125, "completions/mean_terminated_length": 435.125, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.0658, "grad_norm": 0.08944941312074661, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3640459.0, "reward": 0.6743512153625488, "reward_std": 0.6203274726867676, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34869998693466187, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.96812504529953, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 636.3125, "completions/mean_terminated_length": 610.4666748046875, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.066, "grad_norm": 0.06366076320409775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3654864.0, "reward": 0.2871885299682617, "reward_std": 0.5180173516273499, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1368750035762787, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.8431249856948853, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 646.9375, "completions/mean_terminated_length": 621.800048828125, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.0662, "grad_norm": 0.04389781132340431, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3668863.0, "reward": 0.3384258449077606, "reward_std": 0.31520313024520874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11435000598430634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8362499475479126, "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": 0.0, "completions/max_length": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 492.1875, "completions/mean_terminated_length": 492.1875, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 0.0664, "grad_norm": 0.07636898756027222, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3680978.0, "reward": 0.640157163143158, "reward_std": 0.4887438416481018, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4053124785423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6875, "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": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 494.0, "completions/mean_terminated_length": 494.0, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.0666, "grad_norm": 0.06706171482801437, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3694146.0, "reward": 0.2810916602611542, "reward_std": 0.6043267250061035, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1871812492609024, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 638.0, "completions/max_terminated_length": 638.0, "completions/mean_length": 432.9375, "completions/mean_terminated_length": 432.9375, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0668, "grad_norm": 0.07523827999830246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3705057.0, "reward": 0.8276293277740479, "reward_std": 0.6476131081581116, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4677562415599823, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9768750667572021, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 452.0, "completions/mean_terminated_length": 413.86669921875, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.067, "grad_norm": 0.05956811085343361, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3716945.0, "reward": 0.16804784536361694, "reward_std": 0.40844953060150146, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08609375357627869, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.1875, "rewards/mean_confidence_reward": 0.71875, "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": 0.0, "completions/max_length": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 298.5625, "completions/mean_terminated_length": 298.5625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.0672, "grad_norm": 0.11389327049255371, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3725658.0, "reward": 0.8815388083457947, "reward_std": 0.5870842933654785, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45057499408721924, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9818750023841858, "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": 303.0, "completions/max_terminated_length": 303.0, "completions/mean_length": 247.875, "completions/mean_terminated_length": 247.875, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 0.0674, "grad_norm": 0.1566033959388733, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3733672.0, "reward": 0.49953263998031616, "reward_std": 0.706400454044342, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3115624785423279, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7931250333786011, "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": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 396.125, "completions/mean_terminated_length": 396.125, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0676, "grad_norm": 0.09204301238059998, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3744186.0, "reward": 0.4434890151023865, "reward_std": 0.574471116065979, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1994750052690506, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9168750047683716, "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": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 349.4375, "completions/mean_terminated_length": 349.4375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.0678, "grad_norm": 0.143206387758255, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3753969.0, "reward": 0.8275012373924255, "reward_std": 0.64747154712677, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4675000011920929, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.90625, "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": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 454.5, "completions/mean_terminated_length": 454.5, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.068, "grad_norm": 0.040681205689907074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3765273.0, "reward": 0.13476359844207764, "reward_std": 0.24122647941112518, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.01952499896287918, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8993750214576721, "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": 0.0, "completions/max_length": 346.0, "completions/max_terminated_length": 346.0, "completions/mean_length": 252.8125, "completions/mean_terminated_length": 252.8125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.0682, "grad_norm": 0.1372273713350296, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3773622.0, "reward": 0.9131855964660645, "reward_std": 0.6330302357673645, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5138687491416931, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 839.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 513.625, "completions/mean_terminated_length": 513.625, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.0684, "grad_norm": 0.101740263402462, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3785536.0, "reward": 0.5378105640411377, "reward_std": 0.6178989410400391, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2631187438964844, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9099999666213989, "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": 656.0, "completions/max_terminated_length": 656.0, "completions/mean_length": 469.0, "completions/mean_terminated_length": 469.0, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.0686, "grad_norm": 0.08724040538072586, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3796640.0, "reward": 0.6103760004043579, "reward_std": 0.5809032320976257, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28325000405311584, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 454.1875, "completions/mean_terminated_length": 454.1875, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.0688, "grad_norm": 0.10316300392150879, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3807923.0, "reward": 0.6627072691917419, "reward_std": 0.6987124085426331, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3879125118255615, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9706250429153442, "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": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 527.375, "completions/mean_terminated_length": 527.375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.069, "grad_norm": 0.09004779160022736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3820169.0, "reward": 0.4134262502193451, "reward_std": 0.5848204493522644, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20184999704360962, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9106250405311584, "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": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 265.6875, "completions/mean_terminated_length": 265.6875, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.0692, "grad_norm": 0.12133362144231796, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3828276.0, "reward": 0.8262729644775391, "reward_std": 0.5654053092002869, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4025437533855438, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.96875, "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": 0.0, "completions/max_length": 387.0, "completions/max_terminated_length": 387.0, "completions/mean_length": 290.0, "completions/mean_terminated_length": 290.0, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.0694, "grad_norm": 0.11445284634828568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3837492.0, "reward": 0.6498419046401978, "reward_std": 0.6431859135627747, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36218124628067017, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 605.125, "completions/mean_terminated_length": 605.125, "completions/min_length": 386.0, "completions/min_terminated_length": 386.0, "epoch": 0.0696, "grad_norm": 0.03743642568588257, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3851222.0, "reward": 0.1917199343442917, "reward_std": 0.30373284220695496, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07093749940395355, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.3125, "rewards/mean_confidence_reward": 0.815625011920929, "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": 327.0, "completions/max_terminated_length": 327.0, "completions/mean_length": 229.6875, "completions/mean_terminated_length": 229.6875, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.0698, "grad_norm": 0.07051744312047958, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3859025.0, "reward": 0.6134353876113892, "reward_std": 0.40398290753364563, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22686874866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9149999618530273, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 773.8125, "completions/mean_terminated_length": 716.0769653320312, "completions/min_length": 356.0, "completions/min_terminated_length": 356.0, "epoch": 0.07, "grad_norm": 0.04176754504442215, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3876398.0, "reward": 0.2617195248603821, "reward_std": 0.3642440736293793, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1484375, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5843750238418579, "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": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 432.4375, "completions/mean_terminated_length": 432.4375, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.0702, "grad_norm": 0.08231008052825928, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3886981.0, "reward": 0.5774229764938354, "reward_std": 0.6003322005271912, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2798437476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.953125, "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": 0.0, "completions/max_length": 576.0, "completions/max_terminated_length": 576.0, "completions/mean_length": 438.0625, "completions/mean_terminated_length": 438.0625, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.0704, "grad_norm": 0.0657210573554039, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3898326.0, "reward": 0.4183575510978699, "reward_std": 0.4933055341243744, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14921249449253082, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9781249761581421, "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": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 348.5, "completions/mean_terminated_length": 348.5, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.0706, "grad_norm": 0.10313428938388824, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3907262.0, "reward": 0.7300478219985962, "reward_std": 0.6532469987869263, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.39759373664855957, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9031250476837158, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 685.3125, "completions/mean_terminated_length": 636.9285888671875, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.0708, "grad_norm": 0.03608859330415726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3922355.0, "reward": 0.1480482816696167, "reward_std": 0.27359825372695923, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04609374701976776, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.25, "rewards/mean_confidence_reward": 0.8400000333786011, "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": 946.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 456.0, "completions/mean_terminated_length": 456.0, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.071, "grad_norm": 0.04431037977337837, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3933987.0, "reward": 0.40632903575897217, "reward_std": 0.29370561242103577, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1251562535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 382.8125, "completions/mean_terminated_length": 382.8125, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.0712, "grad_norm": 0.09951712936162949, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3943984.0, "reward": 0.37124794721603394, "reward_std": 0.40954217314720154, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11749374866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9493749737739563, "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": 0.0, "completions/max_length": 995.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 640.75, "completions/mean_terminated_length": 640.75, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.0714, "grad_norm": 0.05899391323328018, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3958156.0, "reward": 0.44882893562316895, "reward_std": 0.3879825174808502, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.27265623211860657, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.731249988079071, "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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 330.5625, "completions/mean_terminated_length": 330.5625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.0716, "grad_norm": 0.051327306777238846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3967557.0, "reward": 0.3754572570323944, "reward_std": 0.26153120398521423, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06341250240802765, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 929.125, "completions/mean_terminated_length": 855.3333129882812, "completions/min_length": 690.0, "completions/min_terminated_length": 690.0, "epoch": 0.0718, "grad_norm": 0.015066026709973812, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 3986775.0, "reward": 0.03429735079407692, "reward_std": 0.1371876299381256, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.006093749776482582, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.0625, "rewards/mean_confidence_reward": 0.4156250059604645, "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": 532.0, "completions/max_terminated_length": 532.0, "completions/mean_length": 364.125, "completions/mean_terminated_length": 364.125, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.072, "grad_norm": 0.06166090443730354, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 3996441.0, "reward": 0.562104344367981, "reward_std": 0.41368409991264343, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18670624494552612, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8768750429153442, "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": 0.0, "completions/max_length": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 333.625, "completions/mean_terminated_length": 333.625, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.0722, "grad_norm": 0.06798554211854935, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4006515.0, "reward": 0.2886731028556824, "reward_std": 0.30961838364601135, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07734374701976776, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9375, "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": 0.0, "completions/max_length": 491.0, "completions/max_terminated_length": 491.0, "completions/mean_length": 325.9375, "completions/mean_terminated_length": 325.9375, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.0724, "grad_norm": 0.09941168129444122, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4017250.0, "reward": 0.7021479606628418, "reward_std": 0.5323180556297302, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3417937457561493, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9387500286102295, "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": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 492.4375, "completions/mean_terminated_length": 492.4375, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.0726, "grad_norm": 0.0897354930639267, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4029417.0, "reward": 0.663322925567627, "reward_std": 0.6441425681114197, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.38914376497268677, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9331250190734863, "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": 0.0, "completions/max_length": 904.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 636.9375, "completions/mean_terminated_length": 636.9375, "completions/min_length": 389.0, "completions/min_terminated_length": 389.0, "epoch": 0.0728, "grad_norm": 0.020763885229825974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4043832.0, "reward": 0.06843884289264679, "reward_std": 0.18780949711799622, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.011874999850988388, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 0.0, "completions/max_length": 626.0, "completions/max_terminated_length": 626.0, "completions/mean_length": 510.1875, "completions/mean_terminated_length": 510.1875, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.073, "grad_norm": 0.045094411820173264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4055915.0, "reward": 0.2958357334136963, "reward_std": 0.4156000018119812, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0916687548160553, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 528.0, "completions/max_terminated_length": 528.0, "completions/mean_length": 348.625, "completions/mean_terminated_length": 348.625, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.0732, "grad_norm": 0.0412575863301754, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4066565.0, "reward": 0.36146679520606995, "reward_std": 0.25237607955932617, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03543125092983246, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9756250381469727, "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": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 671.875, "completions/mean_terminated_length": 671.875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.0734, "grad_norm": 0.059506576508283615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4081331.0, "reward": 0.5935041904449463, "reward_std": 0.509221076965332, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24950623512268066, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9268749952316284, "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": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 529.375, "completions/mean_terminated_length": 529.375, "completions/min_length": 370.0, "completions/min_terminated_length": 370.0, "epoch": 0.0736, "grad_norm": 0.08859097212553024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4093609.0, "reward": 0.4461728632450104, "reward_std": 0.6560003757476807, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26734375953674316, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.8031250238418579, "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": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 553.9375, "completions/mean_terminated_length": 553.9375, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 0.0738, "grad_norm": 0.11221768707036972, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4106712.0, "reward": 0.900626003742218, "reward_std": 0.5997474193572998, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5512499809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8468749523162842, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 668.5625, "completions/mean_terminated_length": 550.0833740234375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.074, "grad_norm": 0.043641991913318634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4121217.0, "reward": 0.3864727020263672, "reward_std": 0.38930779695510864, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.21044375002384186, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5606250166893005, "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": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 451.75, "completions/mean_terminated_length": 451.75, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0742, "grad_norm": 0.06491614133119583, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4132317.0, "reward": 0.6929692029953003, "reward_std": 0.3624350428581238, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.573437511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4468750059604645, "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": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 580.625, "completions/mean_terminated_length": 580.625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.0744, "grad_norm": 0.034455060958862305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4145351.0, "reward": 0.2352292835712433, "reward_std": 0.2757309079170227, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03295625001192093, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 514.5625, "completions/mean_terminated_length": 441.7857360839844, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.0746, "grad_norm": 0.0814606323838234, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4158416.0, "reward": 0.3272664546966553, "reward_std": 0.5188472270965576, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15453124046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.765625, "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": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 333.75, "completions/mean_terminated_length": 333.75, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.0748, "grad_norm": 0.11438381671905518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4167676.0, "reward": 0.6620199680328369, "reward_std": 0.6987777948379517, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38653749227523804, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 443.25, "completions/mean_terminated_length": 443.25, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.075, "grad_norm": 0.07342413812875748, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4179488.0, "reward": 0.3366415798664093, "reward_std": 0.41583696007728577, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11078125238418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9487500190734863, "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": 0.0, "completions/max_length": 973.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 609.625, "completions/mean_terminated_length": 609.625, "completions/min_length": 380.0, "completions/min_terminated_length": 380.0, "epoch": 0.0752, "grad_norm": 0.04056399688124657, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4193658.0, "reward": 0.2218637615442276, "reward_std": 0.3021118938922882, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06872499734163284, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.7612500190734863, "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": 865.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 480.125, "completions/mean_terminated_length": 480.125, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.0754, "grad_norm": 0.0919739231467247, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4205644.0, "reward": 0.6913949251174927, "reward_std": 0.6273844838142395, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3827875256538391, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8762500286102295, "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": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 333.125, "completions/mean_terminated_length": 333.125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.0756, "grad_norm": 0.12186700105667114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4215054.0, "reward": 0.9426572918891907, "reward_std": 0.6752617359161377, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5728124976158142, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.949999988079071, "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": 697.0, "completions/max_terminated_length": 697.0, "completions/mean_length": 452.1875, "completions/mean_terminated_length": 452.1875, "completions/min_length": 372.0, "completions/min_terminated_length": 372.0, "epoch": 0.0758, "grad_norm": 0.08405089378356934, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4226337.0, "reward": 0.6246885061264038, "reward_std": 0.6053844094276428, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37437498569488525, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.762499988079071, "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": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 238.5, "completions/mean_terminated_length": 238.5, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.076, "grad_norm": 0.06807811558246613, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4233545.0, "reward": 0.6299142241477966, "reward_std": 0.33986642956733704, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1348249912261963, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9950000047683716, "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": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 253.3125, "completions/mean_terminated_length": 253.3125, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.0762, "grad_norm": 0.10280485451221466, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4242126.0, "reward": 0.78226637840271, "reward_std": 0.46430087089538574, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5020312666893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.653124988079071, "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": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 301.4375, "completions/mean_terminated_length": 301.4375, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.0764, "grad_norm": 0.04208069667220116, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4250997.0, "reward": 0.5728577375411987, "reward_std": 0.24744103848934174, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08321250230073929, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9862499833106995, "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": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 362.875, "completions/mean_terminated_length": 362.875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.0766, "grad_norm": 0.08779662102460861, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4261203.0, "reward": 0.6803136467933655, "reward_std": 0.5315896272659302, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2981249988079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9593749642372131, "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": 691.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 539.3125, "completions/mean_terminated_length": 539.3125, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.0768, "grad_norm": 0.081772081553936, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4273880.0, "reward": 0.5147666931152344, "reward_std": 0.6325962543487549, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.27953124046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9556249976158142, "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": 0.0, "completions/max_length": 565.0, "completions/max_terminated_length": 565.0, "completions/mean_length": 392.75, "completions/mean_terminated_length": 392.75, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.077, "grad_norm": 0.04181218147277832, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4284308.0, "reward": 0.42265719175338745, "reward_std": 0.2542279064655304, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09531249850988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "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": 322.0, "completions/max_terminated_length": 322.0, "completions/mean_length": 233.625, "completions/mean_terminated_length": 233.625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.0772, "grad_norm": 0.15439337491989136, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4292110.0, "reward": 1.046560525894165, "reward_std": 0.5455797910690308, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5931187272071838, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 245.1875, "completions/mean_terminated_length": 245.1875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0774, "grad_norm": 0.14495359361171722, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4299761.0, "reward": 1.0487478971481323, "reward_std": 0.5434747934341431, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5974937677383423, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 269.0, "completions/mean_terminated_length": 269.0, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.0776, "grad_norm": 0.12202541530132294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4308145.0, "reward": 1.0582009553909302, "reward_std": 0.534299910068512, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6164000034332275, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9306249618530273, "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": 0.0, "completions/max_length": 280.0, "completions/max_terminated_length": 280.0, "completions/mean_length": 176.25, "completions/mean_terminated_length": 176.25, "completions/min_length": 93.0, "completions/min_terminated_length": 93.0, "epoch": 0.0778, "grad_norm": 0.06251048296689987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4314597.0, "reward": 0.5257701277732849, "reward_std": 0.1890236884355545, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11403749883174896, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.909375011920929, "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": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 457.125, "completions/mean_terminated_length": 457.125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.078, "grad_norm": 0.10251471400260925, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4325559.0, "reward": 0.7729699611663818, "reward_std": 0.6228036880493164, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4209374785423279, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 734.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 295.6875, "completions/mean_terminated_length": 295.6875, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.0782, "grad_norm": 0.09148888289928436, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4333698.0, "reward": 0.5837292075157166, "reward_std": 0.3998153507709503, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1674562692642212, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9656250476837158, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 798.375, "completions/mean_terminated_length": 723.1666870117188, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.0784, "grad_norm": 0.03192230686545372, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4351496.0, "reward": 0.24687568843364716, "reward_std": 0.34635016322135925, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11874999850988388, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5806249976158142, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 573.375, "completions/mean_terminated_length": 543.3333740234375, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0786, "grad_norm": 0.0487983413040638, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4364878.0, "reward": 0.29820388555526733, "reward_std": 0.3641476631164551, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1589062511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.7431249618530273, "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": 698.0, "completions/max_terminated_length": 698.0, "completions/mean_length": 502.4375, "completions/mean_terminated_length": 502.4375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.0788, "grad_norm": 0.05874287709593773, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4376453.0, "reward": 0.5314853191375732, "reward_std": 0.3486610949039459, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.31296876072883606, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7749999761581421, "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": 0.0, "completions/max_length": 509.0, "completions/max_terminated_length": 509.0, "completions/mean_length": 321.25, "completions/mean_terminated_length": 321.25, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.079, "grad_norm": 0.07760048657655716, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4386105.0, "reward": 0.7095981240272522, "reward_std": 0.5032666325569153, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29419374465942383, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 349.125, "completions/mean_terminated_length": 349.125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.0792, "grad_norm": 0.09471701085567474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4395339.0, "reward": 1.255313754081726, "reward_std": 0.5392025113105774, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8231250047683716, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.909375011920929, "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": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 411.6875, "completions/mean_terminated_length": 411.6875, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.0794, "grad_norm": 0.09200924634933472, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4406118.0, "reward": 1.0488946437835693, "reward_std": 0.5430359840393066, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5977874994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 519.4375, "completions/mean_terminated_length": 519.4375, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.0796, "grad_norm": 0.07778170704841614, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4418509.0, "reward": 0.4707821309566498, "reward_std": 0.3983972370624542, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1915625035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8968750238418579, "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": 0.0, "completions/max_length": 897.0, "completions/max_terminated_length": 897.0, "completions/mean_length": 490.5625, "completions/mean_terminated_length": 490.5625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.0798, "grad_norm": 0.0683170035481453, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4430118.0, "reward": 0.6118603944778442, "reward_std": 0.4495760202407837, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2862187325954437, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7387499809265137, "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": 940.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 569.0, "completions/mean_terminated_length": 569.0, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.08, "grad_norm": 0.07019452750682831, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4443558.0, "reward": 0.31742241978645325, "reward_std": 0.4407772123813629, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.25984376668930054, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.2875000238418579, "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": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 388.0, "completions/mean_terminated_length": 388.0, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.0802, "grad_norm": 0.10469542443752289, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4454038.0, "reward": 0.4705171585083008, "reward_std": 0.6446486711502075, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.25353124737739563, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9912500381469727, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 727.0625, "completions/mean_terminated_length": 658.5385131835938, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.0804, "grad_norm": 0.03272178769111633, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4469351.0, "reward": 0.3218727707862854, "reward_std": 0.29832950234413147, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08124375343322754, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.6962499618530273, "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": 0.0, "completions/max_length": 309.0, "completions/max_terminated_length": 309.0, "completions/mean_length": 250.5625, "completions/mean_terminated_length": 250.5625, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.0806, "grad_norm": 0.13078002631664276, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4476976.0, "reward": 0.8280417919158936, "reward_std": 0.5645126104354858, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40608125925064087, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 297.0, "completions/mean_terminated_length": 297.0, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.0808, "grad_norm": 0.09230512380599976, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4485024.0, "reward": 0.8973103761672974, "reward_std": 0.4818549156188965, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41961875557899475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 467.1875, "completions/mean_terminated_length": 467.1875, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.081, "grad_norm": 0.09984754770994186, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4496595.0, "reward": 0.9088791608810425, "reward_std": 0.7101213335990906, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5677562355995178, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.921875, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 660.625, "completions/mean_terminated_length": 636.4000244140625, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.0812, "grad_norm": 0.058799635618925095, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4511629.0, "reward": 0.7549222707748413, "reward_std": 0.4605385959148407, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6348437666893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.27812498807907104, "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": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 356.5625, "completions/mean_terminated_length": 356.5625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.0814, "grad_norm": 0.10357989370822906, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4521110.0, "reward": 1.0098166465759277, "reward_std": 0.596150815486908, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5821312665939331, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9031250476837158, "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": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 429.375, "completions/mean_terminated_length": 429.375, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.0816, "grad_norm": 0.06407195329666138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4532668.0, "reward": 0.5671882629394531, "reward_std": 0.3746801018714905, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3843750059604645, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.640625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 592.8125, "completions/mean_terminated_length": 564.0667114257812, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0818, "grad_norm": 0.06416510790586472, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4546297.0, "reward": 0.4770318865776062, "reward_std": 0.42495787143707275, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.26656249165534973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5625, "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": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 490.875, "completions/mean_terminated_length": 490.875, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.082, "grad_norm": 0.083651602268219, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4558167.0, "reward": 0.6974886655807495, "reward_std": 0.6509780883789062, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4574749767780304, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7987500429153442, "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": 763.0, "completions/max_terminated_length": 763.0, "completions/mean_length": 592.0, "completions/mean_terminated_length": 592.0, "completions/min_length": 410.0, "completions/min_terminated_length": 410.0, "epoch": 0.0822, "grad_norm": 0.05131648853421211, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4572503.0, "reward": 0.4516414701938629, "reward_std": 0.42543256282806396, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.21578124165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8456249833106995, "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": 352.0, "completions/max_terminated_length": 352.0, "completions/mean_length": 212.125, "completions/mean_terminated_length": 212.125, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.0824, "grad_norm": 0.11406425386667252, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4579705.0, "reward": 1.3462855815887451, "reward_std": 0.4302581548690796, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8800687789916992, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9725000262260437, "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": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 438.375, "completions/mean_terminated_length": 438.375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.0826, "grad_norm": 0.10848947614431381, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4590783.0, "reward": 0.6173980236053467, "reward_std": 0.6623812913894653, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3597937822341919, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9506250023841858, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 776.875, "completions/mean_terminated_length": 741.5714721679688, "completions/min_length": 466.0, "completions/min_terminated_length": 466.0, "epoch": 0.0828, "grad_norm": 0.029504796490073204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4607069.0, "reward": 0.08148512989282608, "reward_std": 0.22581946849822998, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.037968747317790985, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.5687500238418579, "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": 426.0, "completions/max_terminated_length": 426.0, "completions/mean_length": 305.5, "completions/mean_terminated_length": 305.5, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.083, "grad_norm": 0.1568153351545334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4615765.0, "reward": 0.9462916851043701, "reward_std": 0.6737523078918457, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5800812244415283, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 390.75, "completions/mean_terminated_length": 390.75, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.0832, "grad_norm": 0.09271752834320068, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4626145.0, "reward": 0.6624103784561157, "reward_std": 0.5525385737419128, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32481876015663147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8574999570846558, "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": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 578.875, "completions/mean_terminated_length": 578.875, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.0834, "grad_norm": 0.054214879870414734, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4639407.0, "reward": 0.42779475450515747, "reward_std": 0.49688422679901123, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16808749735355377, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8968750238418579, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 882.0, "completions/mean_length": 771.3125, "completions/mean_terminated_length": 656.45458984375, "completions/min_length": 391.0, "completions/min_terminated_length": 391.0, "epoch": 0.0836, "grad_norm": 0.04869672283530235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4655652.0, "reward": 0.3350007236003876, "reward_std": 0.5240011811256409, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17000000178813934, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.484375, "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": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 383.375, "completions/mean_terminated_length": 383.375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.0838, "grad_norm": 0.08552606403827667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4665498.0, "reward": 1.1334134340286255, "reward_std": 0.581059992313385, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7043249607086182, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 898.0, "completions/mean_length": 654.0, "completions/mean_terminated_length": 629.3333740234375, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.084, "grad_norm": 0.05325153097510338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4680650.0, "reward": 0.5750786662101746, "reward_std": 0.4232742488384247, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46265625953674316, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.42499998211860657, "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": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 433.6875, "completions/mean_terminated_length": 433.6875, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 0.0842, "grad_norm": 0.11707399785518646, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4691605.0, "reward": 0.6902886629104614, "reward_std": 0.7489873766899109, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4430750012397766, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9762499928474426, "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": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 446.1875, "completions/mean_terminated_length": 446.1875, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 0.0844, "grad_norm": 0.06174736097455025, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4703176.0, "reward": 0.41679802536964417, "reward_std": 0.30469566583633423, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14609374105930328, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8968750238418579, "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": 0.0, "completions/max_length": 1003.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 438.125, "completions/mean_terminated_length": 438.125, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0846, "grad_norm": 0.0591457299888134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4714234.0, "reward": 0.4628794193267822, "reward_std": 0.3650687634944916, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11325625330209732, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 318.75, "completions/mean_terminated_length": 318.75, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.0848, "grad_norm": 0.0621362030506134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4723110.0, "reward": 0.6495291590690613, "reward_std": 0.4204282760620117, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2990562617778778, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8306249976158142, "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": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 568.3125, "completions/mean_terminated_length": 568.3125, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.085, "grad_norm": 0.04970346763730049, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4737067.0, "reward": 0.5285040140151978, "reward_std": 0.44749563932418823, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18200625479221344, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 0.0, "completions/max_length": 388.0, "completions/max_terminated_length": 388.0, "completions/mean_length": 341.0, "completions/mean_terminated_length": 341.0, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.0852, "grad_norm": 0.11865879595279694, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4746683.0, "reward": 0.6045541763305664, "reward_std": 0.5815684795379639, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2716062664985657, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 252.0, "completions/max_terminated_length": 252.0, "completions/mean_length": 176.0625, "completions/mean_terminated_length": 176.0625, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 0.0854, "grad_norm": 0.23438116908073425, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4753308.0, "reward": 0.8838983178138733, "reward_std": 0.6658615469932556, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5177937746047974, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.984375, "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": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 492.125, "completions/mean_terminated_length": 492.125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.0856, "grad_norm": 0.07380641251802444, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4765262.0, "reward": 0.6535947322845459, "reward_std": 0.516930103302002, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3696874976158142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7906249761581421, "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": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 429.0, "completions/mean_terminated_length": 429.0, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.0858, "grad_norm": 0.09015947580337524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4776046.0, "reward": 0.8644509315490723, "reward_std": 0.6126791834831238, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4788999855518341, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9493750333786011, "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": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 337.6875, "completions/mean_terminated_length": 337.6875, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.086, "grad_norm": 0.1143970862030983, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4785641.0, "reward": 1.1311947107315063, "reward_std": 0.5838762521743774, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6998875141143799, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 765.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 467.9375, "completions/mean_terminated_length": 467.9375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.0862, "grad_norm": 0.022022519260644913, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4799240.0, "reward": 0.48015719652175903, "reward_std": 0.18743380904197693, "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": 0.949999988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 531.1875, "completions/mean_terminated_length": 498.3333740234375, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.0864, "grad_norm": 0.06629869341850281, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4811803.0, "reward": 0.8148196339607239, "reward_std": 0.6149912476539612, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5046374797821045, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7806249856948853, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 649.1875, "completions/mean_terminated_length": 595.6428833007812, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.0866, "grad_norm": 0.06258974224328995, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4827134.0, "reward": 0.18749824166297913, "reward_std": 0.5123390555381775, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.12499374896287918, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.125, "rewards/mean_confidence_reward": 0.8650000095367432, "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": 960.0, "completions/max_terminated_length": 960.0, "completions/mean_length": 499.25, "completions/mean_terminated_length": 499.25, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.0868, "grad_norm": 0.08451417833566666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4839090.0, "reward": 0.684219241142273, "reward_std": 0.5033432841300964, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6184375286102295, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.3375000059604645, "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": 970.0, "completions/mean_length": 638.1875, "completions/mean_terminated_length": 612.4666748046875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.087, "grad_norm": 0.11260061711072922, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4853157.0, "reward": 0.6678134202957153, "reward_std": 0.6975295543670654, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3981249928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.8893749713897705, "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": 613.0, "completions/max_terminated_length": 613.0, "completions/mean_length": 347.0625, "completions/mean_terminated_length": 347.0625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.0872, "grad_norm": 0.11712773889303207, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4863526.0, "reward": 0.4501509666442871, "reward_std": 0.5753114223480225, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2128000110387802, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9575000405311584, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 609.5, "completions/mean_terminated_length": 581.86669921875, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.0874, "grad_norm": 0.06432927399873734, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4877662.0, "reward": 0.3989076614379883, "reward_std": 0.3972921073436737, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11031249910593033, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9093750715255737, "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": 255.0, "completions/max_terminated_length": 255.0, "completions/mean_length": 186.8125, "completions/mean_terminated_length": 186.8125, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.0876, "grad_norm": 0.16962428390979767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4884635.0, "reward": 1.3120886087417603, "reward_std": 0.5121867060661316, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.874174952507019, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9768750667572021, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 707.375, "completions/mean_terminated_length": 662.1428833007812, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.0878, "grad_norm": 0.057845309376716614, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4900353.0, "reward": 0.5973440408706665, "reward_std": 0.48019376397132874, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5696874856948853, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.2250000238418579, "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": 0.0, "completions/max_length": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 438.75, "completions/mean_terminated_length": 438.75, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.088, "grad_norm": 0.12583371996879578, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4912109.0, "reward": 0.6791414022445679, "reward_std": 0.6741499900817871, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4832812547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.7250000238418579, "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": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 640.0625, "completions/mean_terminated_length": 640.0625, "completions/min_length": 506.0, "completions/min_terminated_length": 506.0, "epoch": 0.0882, "grad_norm": 0.07487773150205612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4926462.0, "reward": 0.756460428237915, "reward_std": 0.7044768333435059, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45041877031326294, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9481250047683716, "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": 369.0, "completions/max_terminated_length": 369.0, "completions/mean_length": 291.0625, "completions/mean_terminated_length": 291.0625, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.0884, "grad_norm": 0.12024299800395966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4935135.0, "reward": 1.3116166591644287, "reward_std": 0.5120023488998413, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8732312321662903, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 658.0, "completions/max_terminated_length": 658.0, "completions/mean_length": 506.875, "completions/mean_terminated_length": 506.875, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.0886, "grad_norm": 0.08347392082214355, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4947453.0, "reward": 0.4122545123100281, "reward_std": 0.5847313404083252, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.1995062530040741, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9143750667572021, "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": 861.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 550.25, "completions/mean_terminated_length": 550.25, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.0888, "grad_norm": 0.08818341046571732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4960401.0, "reward": 0.4355476200580597, "reward_std": 0.4584142565727234, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.37109375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.4937500059604645, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 774.875, "completions/mean_terminated_length": 758.2667236328125, "completions/min_length": 441.0, "completions/min_terminated_length": 441.0, "epoch": 0.089, "grad_norm": 0.058112505823373795, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4976639.0, "reward": 0.35015714168548584, "reward_std": 0.5073099732398987, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2628124952316284, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.375, "rewards/mean_confidence_reward": 0.5806249976158142, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 574.0625, "completions/mean_terminated_length": 544.0667114257812, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.0892, "grad_norm": 0.08588546514511108, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 4989904.0, "reward": 0.38343822956085205, "reward_std": 0.5018161535263062, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2668749988079071, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.49312496185302734, "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": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 373.5625, "completions/mean_terminated_length": 373.5625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.0894, "grad_norm": 0.1016162782907486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 4999529.0, "reward": 0.8022637367248535, "reward_std": 0.5029736161231995, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35452502965927124, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9649999737739563, "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": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 314.0, "completions/mean_terminated_length": 314.0, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.0896, "grad_norm": 0.05212130397558212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5007977.0, "reward": 0.6471883058547974, "reward_std": 0.23705986142158508, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3568750023841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7000000476837158, "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": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 418.625, "completions/mean_terminated_length": 418.625, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.0898, "grad_norm": 0.09376665949821472, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5018531.0, "reward": 0.577551007270813, "reward_std": 0.6004311442375183, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28009992837905884, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.8949375152587891, "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": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 621.5625, "completions/mean_terminated_length": 621.5625, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.09, "grad_norm": 0.07104604691267014, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5032540.0, "reward": 0.5774887204170227, "reward_std": 0.6004394888877869, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2799749970436096, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 431.875, "completions/mean_terminated_length": 431.875, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.0902, "grad_norm": 0.04535903036594391, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5043370.0, "reward": 0.4657040238380432, "reward_std": 0.29042381048202515, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18140624463558197, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.824999988079071, "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": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 476.125, "completions/mean_terminated_length": 476.125, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.0904, "grad_norm": 0.04291992262005806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5057884.0, "reward": 0.3651292026042938, "reward_std": 0.25483280420303345, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04275624826550484, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9493750333786011, "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": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 447.5625, "completions/mean_terminated_length": 447.5625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.0906, "grad_norm": 0.052000273019075394, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5068773.0, "reward": 0.35101670026779175, "reward_std": 0.28318649530410767, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07703125476837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9456249475479126, "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": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 399.0, "completions/mean_terminated_length": 399.0, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.0908, "grad_norm": 0.04665278270840645, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5078965.0, "reward": 0.3772665858268738, "reward_std": 0.26269403100013733, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06703124940395355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.890625, "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": 527.0, "completions/max_terminated_length": 527.0, "completions/mean_length": 413.1875, "completions/mean_terminated_length": 413.1875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.091, "grad_norm": 0.13986298441886902, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5089640.0, "reward": 0.8433544635772705, "reward_std": 0.768159806728363, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5617062449455261, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9793750047683716, "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": 903.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 424.8125, "completions/mean_terminated_length": 424.8125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.0912, "grad_norm": 0.10047484934329987, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5100693.0, "reward": 0.9552322626113892, "reward_std": 0.5877017378807068, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.535462498664856, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9556249976158142, "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": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 426.0, "completions/mean_terminated_length": 426.0, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.0914, "grad_norm": 0.057879865169525146, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5111589.0, "reward": 0.4017009437084198, "reward_std": 0.30213332176208496, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11589999496936798, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7756249904632568, "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": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 448.0625, "completions/mean_terminated_length": 448.0625, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.0916, "grad_norm": 0.06576050072908401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5122950.0, "reward": 0.4161728620529175, "reward_std": 0.40442168712615967, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14484375715255737, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.9156250357627869, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 421.6875, "completions/mean_terminated_length": 335.64288330078125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.0918, "grad_norm": 0.13414613902568817, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5133425.0, "reward": 0.7523446083068848, "reward_std": 0.6795708537101746, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5046875476837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6656249761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 379.375, "completions/mean_terminated_length": 336.4000244140625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.092, "grad_norm": 0.1024683490395546, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5144263.0, "reward": 0.9442980289459229, "reward_std": 0.5312497019767761, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5135937333106995, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.871874988079071, "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": 439.0, "completions/max_terminated_length": 439.0, "completions/mean_length": 184.75, "completions/mean_terminated_length": 184.75, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 0.0922, "grad_norm": 0.16479331254959106, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5150899.0, "reward": 1.2806107997894287, "reward_std": 0.5154231786727905, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8737187385559082, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8668749928474426, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 457.9375, "completions/mean_terminated_length": 420.20001220703125, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.0924, "grad_norm": 0.06278634071350098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5162194.0, "reward": 0.55066978931427, "reward_std": 0.4255889356136322, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1638375073671341, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9081250429153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 654.0, "completions/mean_length": 386.625, "completions/mean_terminated_length": 344.13336181640625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.0926, "grad_norm": 0.12146230041980743, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5172892.0, "reward": 1.003582239151001, "reward_std": 0.6793515682220459, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6321625113487244, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.8987499475479126, "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": 964.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 484.25, "completions/mean_terminated_length": 484.25, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.0928, "grad_norm": 0.14266566932201385, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5184832.0, "reward": 0.6893758177757263, "reward_std": 0.7479397654533386, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4412500262260437, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.6593749523162842, "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": 0.0, "completions/max_length": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 303.25, "completions/mean_terminated_length": 303.25, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.093, "grad_norm": 0.06820833683013916, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5193588.0, "reward": 0.7288821935653687, "reward_std": 0.3828071355819702, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2702624797821045, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9474999904632568, "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": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 384.5625, "completions/mean_terminated_length": 384.5625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.0932, "grad_norm": 0.09108574688434601, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5203869.0, "reward": 0.7602227926254272, "reward_std": 0.5362550616264343, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.45794373750686646, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8050000071525574, "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": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 229.8125, "completions/mean_terminated_length": 229.8125, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.0934, "grad_norm": 9.029849752550945e-05, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5211130.0, "reward": 1.4998295307159424, "reward_std": 0.0004217267851345241, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9996562600135803, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9906250238418579, "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": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 355.4375, "completions/mean_terminated_length": 355.4375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.0936, "grad_norm": 0.11550825834274292, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5221649.0, "reward": 0.8271887302398682, "reward_std": 0.6471262574195862, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46687498688697815, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 300.4375, "completions/mean_terminated_length": 300.4375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.0938, "grad_norm": 0.08354660868644714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5230376.0, "reward": 0.7808541655540466, "reward_std": 0.42850229144096375, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3117062449455261, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 317.125, "completions/mean_terminated_length": 270.0, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.094, "grad_norm": 0.15947724878787994, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5239226.0, "reward": 0.6952947974205017, "reward_std": 0.5994174480438232, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3280875086784363, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9143750071525574, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 720.1875, "completions/mean_terminated_length": 699.933349609375, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.0942, "grad_norm": 0.07253949344158173, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5255405.0, "reward": 0.6372661590576172, "reward_std": 0.6152471303939819, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5245312452316284, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5218750238418579, "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": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 275.8125, "completions/mean_terminated_length": 275.8125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.0944, "grad_norm": 0.10634859651327133, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5264378.0, "reward": 1.2646448612213135, "reward_std": 0.42010974884033203, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7792875170707703, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 0.0, "completions/max_length": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 252.5625, "completions/mean_terminated_length": 252.5625, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.0946, "grad_norm": 0.09528715908527374, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5272179.0, "reward": 0.6161606311798096, "reward_std": 0.36933648586273193, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.16981874406337738, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 0.0, "completions/max_length": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 373.25, "completions/mean_terminated_length": 373.25, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.0948, "grad_norm": 0.12133172899484634, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5281991.0, "reward": 0.9211447834968567, "reward_std": 0.6278639435768127, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.529787540435791, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 289.0, "completions/mean_length": 255.875, "completions/mean_terminated_length": 204.6666717529297, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.095, "grad_norm": 0.1611882597208023, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5290085.0, "reward": 1.1502294540405273, "reward_std": 0.5390401482582092, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8004562258720398, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7887500524520874, "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": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 470.1875, "completions/mean_terminated_length": 470.1875, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 0.0952, "grad_norm": 0.12829537689685822, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5301944.0, "reward": 0.5342140197753906, "reward_std": 0.6943077445030212, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3184249997138977, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.9831249713897705, "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": 942.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 479.375, "completions/mean_terminated_length": 479.375, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 0.0954, "grad_norm": 0.05201731622219086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5314750.0, "reward": 0.5780477523803711, "reward_std": 0.3385320007801056, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2185937464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8812499642372131, "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": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 333.5, "completions/mean_terminated_length": 333.5, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.0956, "grad_norm": 0.0950482040643692, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5324742.0, "reward": 0.8034759759902954, "reward_std": 0.5019620656967163, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35694998502731323, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 752.25, "completions/mean_terminated_length": 734.1333618164062, "completions/min_length": 486.0, "completions/min_terminated_length": 486.0, "epoch": 0.0958, "grad_norm": 0.028225261718034744, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5341194.0, "reward": 0.4900665283203125, "reward_std": 0.2639633119106293, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16763123869895935, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8324999809265137, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 596.9375, "completions/mean_terminated_length": 535.9285888671875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.096, "grad_norm": 0.07683472335338593, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5354473.0, "reward": 0.6014041900634766, "reward_std": 0.6091201901435852, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3278062641620636, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.7993749976158142, "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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 331.0, "completions/mean_terminated_length": 331.0, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.0962, "grad_norm": 0.07986898720264435, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5363385.0, "reward": 0.6551946401596069, "reward_std": 0.45722246170043945, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2478875070810318, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9393749833106995, "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": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 476.9375, "completions/mean_terminated_length": 476.9375, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.0964, "grad_norm": 0.11766506731510162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5375880.0, "reward": 0.6260074973106384, "reward_std": 0.718264639377594, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3770124912261963, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 473.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 310.125, "completions/mean_terminated_length": 310.125, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.0966, "grad_norm": 0.11021550744771957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5385098.0, "reward": 1.2841228246688843, "reward_std": 0.4753551781177521, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8182437419891357, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 405.375, "completions/mean_terminated_length": 405.375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.0968, "grad_norm": 0.05875449627637863, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5395152.0, "reward": 0.6441791653633118, "reward_std": 0.36877813935279846, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2258562445640564, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9375, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 603.375, "completions/mean_terminated_length": 575.3333740234375, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.097, "grad_norm": 0.09149405360221863, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5409014.0, "reward": 0.62211012840271, "reward_std": 0.5370094180107117, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.36921873688697815, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7962499856948853, "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": 0.0, "completions/max_length": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 517.4375, "completions/mean_terminated_length": 517.4375, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.0972, "grad_norm": 0.11162137985229492, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5422365.0, "reward": 0.739448070526123, "reward_std": 0.6505189538002014, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4163937568664551, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.8987499475479126, "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": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 312.125, "completions/mean_terminated_length": 312.125, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.0974, "grad_norm": 0.13063006103038788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5431295.0, "reward": 0.9553760290145874, "reward_std": 0.5878437161445618, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5357499718666077, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 472.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 339.6875, "completions/mean_terminated_length": 339.6875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.0976, "grad_norm": 0.11766309291124344, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5440938.0, "reward": 1.0360572338104248, "reward_std": 0.5860582590103149, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6346122622489929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9242500066757202, "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": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 257.5, "completions/mean_terminated_length": 257.5, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.0978, "grad_norm": 0.11431456357240677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5448754.0, "reward": 1.0113637447357178, "reward_std": 0.5043516755104065, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.522724986076355, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 725.0625, "completions/mean_terminated_length": 656.0769653320312, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.098, "grad_norm": 0.09229521453380585, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5464451.0, "reward": 0.5775008201599121, "reward_std": 0.652779221534729, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4049999713897705, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.5375000238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 483.9375, "completions/mean_terminated_length": 447.933349609375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.0982, "grad_norm": 0.048243965953588486, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5476610.0, "reward": 0.4813290238380432, "reward_std": 0.25239622592926025, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15015622973442078, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8062499761581421, "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": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 416.4375, "completions/mean_terminated_length": 416.4375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.0984, "grad_norm": 0.12446863949298859, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5487049.0, "reward": 0.8739072680473328, "reward_std": 0.5210626125335693, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5603125095367432, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7625000476837158, "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": 715.0, "completions/max_terminated_length": 715.0, "completions/mean_length": 508.3125, "completions/mean_terminated_length": 508.3125, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.0986, "grad_norm": 0.01838442124426365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5499518.0, "reward": 0.5168761014938354, "reward_std": 0.1413889229297638, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09624999761581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468750357627869, "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": 890.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 356.125, "completions/mean_terminated_length": 356.125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.0988, "grad_norm": 0.08033335953950882, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5509328.0, "reward": 0.6796758770942688, "reward_std": 0.3775605857372284, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2968499958515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8706250190734863, "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": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 579.3125, "completions/mean_terminated_length": 579.3125, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.099, "grad_norm": 0.11825058609247208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5522965.0, "reward": 0.7818354964256287, "reward_std": 0.7517982721328735, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5011687278747559, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9856250286102295, "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": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 329.5625, "completions/mean_terminated_length": 329.5625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.0992, "grad_norm": 0.08197663724422455, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5532142.0, "reward": 1.3455352783203125, "reward_std": 0.42997410893440247, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8785687685012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9487500190734863, "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": 521.0, "completions/max_terminated_length": 521.0, "completions/mean_length": 444.0625, "completions/mean_terminated_length": 444.0625, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.0994, "grad_norm": 0.10039473325014114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5543695.0, "reward": 0.9478667974472046, "reward_std": 0.5942721962928772, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5207312107086182, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9837499856948853, "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": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 337.125, "completions/mean_terminated_length": 337.125, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.0996, "grad_norm": 0.1228455975651741, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5552945.0, "reward": 0.9835917949676514, "reward_std": 0.5481807589530945, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5296812057495117, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 0.0, "completions/max_length": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 464.6875, "completions/mean_terminated_length": 464.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.0998, "grad_norm": 0.07447785884141922, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5563836.0, "reward": 0.6908446550369263, "reward_std": 0.4233750104904175, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2566874921321869, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9543750286102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 596.125, "completions/mean_terminated_length": 567.6000366210938, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.1, "grad_norm": 0.05446093901991844, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5577774.0, "reward": 0.3993760645389557, "reward_std": 0.30077752470970154, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11124999821186066, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.765625, "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": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 520.3125, "completions/mean_terminated_length": 520.3125, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.1002, "grad_norm": 0.06554189324378967, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5590275.0, "reward": 0.9950759410858154, "reward_std": 0.5066954493522644, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.615149974822998, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8243750333786011, "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": 441.0, "completions/max_terminated_length": 441.0, "completions/mean_length": 299.3125, "completions/mean_terminated_length": 299.3125, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.1004, "grad_norm": 0.05846220999956131, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5599064.0, "reward": 0.5194481611251831, "reward_std": 0.31684988737106323, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10139375180006027, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9112499952316284, "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": 838.0, "completions/max_terminated_length": 838.0, "completions/mean_length": 359.9375, "completions/mean_terminated_length": 359.9375, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.1006, "grad_norm": 0.12912055850028992, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5608615.0, "reward": 0.7347513437271118, "reward_std": 0.5700716376304626, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.34450000524520874, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 653.4375, "completions/mean_terminated_length": 628.7333374023438, "completions/min_length": 345.0, "completions/min_terminated_length": 345.0, "epoch": 0.1008, "grad_norm": 0.07424181699752808, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5623246.0, "reward": 0.5823414325714111, "reward_std": 0.467263787984848, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28968125581741333, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7337499856948853, "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": 496.0, "completions/max_terminated_length": 496.0, "completions/mean_length": 339.4375, "completions/mean_terminated_length": 339.4375, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.101, "grad_norm": 0.11570020020008087, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5632405.0, "reward": 0.7654354572296143, "reward_std": 0.4742102026939392, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3433687388896942, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9199999570846558, "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": 0.0, "completions/max_length": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 698.5, "completions/mean_terminated_length": 698.5, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.1012, "grad_norm": 0.025463635101914406, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5649053.0, "reward": 0.842031717300415, "reward_std": 0.26219722628593445, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7465624809265137, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3812499940395355, "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": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 381.9375, "completions/mean_terminated_length": 381.9375, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.1014, "grad_norm": 0.12747706472873688, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5659196.0, "reward": 0.7006856203079224, "reward_std": 0.5071107745170593, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.27636876702308655, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.971875011920929, "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": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 530.5625, "completions/mean_terminated_length": 530.5625, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.1016, "grad_norm": 0.08078381419181824, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5671541.0, "reward": 0.7413290739059448, "reward_std": 0.45591893792152405, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.42015624046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8218749761581421, "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": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 442.625, "completions/mean_terminated_length": 442.625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1018, "grad_norm": 0.14220117032527924, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5682511.0, "reward": 0.658982515335083, "reward_std": 0.6994225978851318, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38046249747276306, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9706250429153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 428.8125, "completions/mean_terminated_length": 389.13336181640625, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.102, "grad_norm": 0.07531654834747314, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5693244.0, "reward": 1.296563744544983, "reward_std": 0.4577373266220093, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8431249856948853, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8937499523162842, "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": 293.0, "completions/max_terminated_length": 293.0, "completions/mean_length": 248.75, "completions/mean_terminated_length": 248.75, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.1022, "grad_norm": 0.05533996596932411, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5701352.0, "reward": 1.440176010131836, "reward_std": 0.23771391808986664, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9428499937057495, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9762499928474426, "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": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 304.125, "completions/mean_terminated_length": 304.125, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.1024, "grad_norm": 0.06292913854122162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5710010.0, "reward": 0.5364853739738464, "reward_std": 0.3167251944541931, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1354687511920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9406249523162842, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 385.375, "completions/mean_terminated_length": 385.375, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.1026, "grad_norm": 0.08769503980875015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5720224.0, "reward": 1.1405197381973267, "reward_std": 0.4786759614944458, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.656037449836731, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674375057220459, "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": 1023.0, "completions/mean_length": 802.8125, "completions/mean_terminated_length": 788.0667114257812, "completions/min_length": 563.0, "completions/min_terminated_length": 563.0, "epoch": 0.1028, "grad_norm": 0.0636250451207161, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5737341.0, "reward": 0.5556010603904724, "reward_std": 0.5303301811218262, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2361999899148941, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9037500023841858, "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": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 277.5625, "completions/mean_terminated_length": 277.5625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.103, "grad_norm": 0.0680822879076004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5746214.0, "reward": 1.4054948091506958, "reward_std": 0.3747987449169159, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9359874725341797, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 428.0, "completions/max_terminated_length": 428.0, "completions/mean_length": 324.9375, "completions/mean_terminated_length": 324.9375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1032, "grad_norm": 0.10815872997045517, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5755429.0, "reward": 1.1660792827606201, "reward_std": 0.5254442095756531, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.707156240940094, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 376.5, "completions/mean_terminated_length": 376.5, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.1034, "grad_norm": 0.10596824437379837, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5764845.0, "reward": 0.8134385347366333, "reward_std": 0.49763795733451843, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37687501311302185, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 0.0, "completions/max_length": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 490.875, "completions/mean_terminated_length": 490.875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.1036, "grad_norm": 0.08210188150405884, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5776427.0, "reward": 1.0641164779663086, "reward_std": 0.5004977583885193, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6907312273979187, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8287500143051147, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 605.75, "completions/mean_terminated_length": 577.86669921875, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.1038, "grad_norm": 0.06559213250875473, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5790263.0, "reward": 0.4803481101989746, "reward_std": 0.46552756428718567, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.14819374680519104, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9162499904632568, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 639.125, "completions/mean_terminated_length": 613.4666748046875, "completions/min_length": 480.0, "completions/min_terminated_length": 480.0, "epoch": 0.104, "grad_norm": 0.04584847390651703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5804329.0, "reward": 0.5616416931152344, "reward_std": 0.4250863194465637, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18578124046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 790.8125, "completions/mean_terminated_length": 713.0833740234375, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.1042, "grad_norm": 0.05461398512125015, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5821670.0, "reward": 0.5222537517547607, "reward_std": 0.5007611513137817, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2945062518119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5674999952316284, "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": 568.0, "completions/max_terminated_length": 568.0, "completions/mean_length": 404.75, "completions/mean_terminated_length": 404.75, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1044, "grad_norm": 0.1051785945892334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5832178.0, "reward": 0.973660409450531, "reward_std": 0.5800950527191162, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5723187327384949, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9293749928474426, "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": 598.0, "completions/max_terminated_length": 598.0, "completions/mean_length": 400.625, "completions/mean_terminated_length": 400.625, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.1046, "grad_norm": 0.06559184938669205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5842860.0, "reward": 0.7573416233062744, "reward_std": 0.3773464560508728, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.32718124985694885, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9181250333786011, "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": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 548.0, "completions/mean_terminated_length": 548.0, "completions/min_length": 430.0, "completions/min_terminated_length": 430.0, "epoch": 0.1048, "grad_norm": 0.04676944389939308, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5855852.0, "reward": 0.30262911319732666, "reward_std": 0.27590471506118774, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04275624826550484, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.9612500667572021, "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": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 484.6875, "completions/mean_terminated_length": 484.6875, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.105, "grad_norm": 0.04331386089324951, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5867271.0, "reward": 0.6484383344650269, "reward_std": 0.2987295091152191, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.421875, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.684374988079071, "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": 721.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 494.8125, "completions/mean_terminated_length": 494.8125, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.1052, "grad_norm": 0.05527740716934204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5879764.0, "reward": 0.6257040500640869, "reward_std": 0.2990116775035858, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2514062523841858, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8843749761581421, "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": 969.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 604.1875, "completions/mean_terminated_length": 604.1875, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.1054, "grad_norm": 0.06525679677724838, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5893623.0, "reward": 0.6864818334579468, "reward_std": 0.391683965921402, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.560462474822998, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.39937499165534973, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 442.75, "completions/mean_terminated_length": 404.0000305175781, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.1056, "grad_norm": 0.06601214408874512, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5904707.0, "reward": 0.522657036781311, "reward_std": 0.3549363911151886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2953124940395355, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.625, "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": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 477.25, "completions/mean_terminated_length": 477.25, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.1058, "grad_norm": 0.09943190217018127, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5916007.0, "reward": 1.2761541604995728, "reward_std": 0.47853201627731323, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8023062348365784, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9412499666213989, "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": 986.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 676.1875, "completions/mean_terminated_length": 676.1875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.106, "grad_norm": 0.09222489595413208, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5931914.0, "reward": 0.6112512350082397, "reward_std": 0.48589181900024414, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2224999964237213, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 408.0, "completions/max_terminated_length": 408.0, "completions/mean_length": 336.6875, "completions/mean_terminated_length": 336.6875, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.1062, "grad_norm": 0.1019909456372261, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 5941333.0, "reward": 1.07606041431427, "reward_std": 0.49561429023742676, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5896187424659729, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 439.6875, "completions/mean_terminated_length": 439.6875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1064, "grad_norm": 0.1435832530260086, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5952128.0, "reward": 0.8104637861251831, "reward_std": 0.6296693682670593, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5584249496459961, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7174999713897705, "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": 650.0, "completions/max_terminated_length": 650.0, "completions/mean_length": 504.4375, "completions/mean_terminated_length": 504.4375, "completions/min_length": 378.0, "completions/min_terminated_length": 378.0, "epoch": 0.1066, "grad_norm": 0.04408768564462662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5964167.0, "reward": 0.5914480686187744, "reward_std": 0.3046373426914215, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1828937530517578, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9175000190734863, "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": 0.0, "completions/max_length": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 411.625, "completions/mean_terminated_length": 411.625, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.1068, "grad_norm": 0.09144098311662674, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5974753.0, "reward": 0.8059353828430176, "reward_std": 0.5007718205451965, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3618687391281128, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9149999618530273, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 774.0, "completions/mean_length": 533.5625, "completions/mean_terminated_length": 463.5000305175781, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.107, "grad_norm": 0.1291322410106659, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 5987642.0, "reward": 0.8181260228157043, "reward_std": 0.7243622541427612, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5112500190734863, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.721875011920929, "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": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 560.5, "completions/mean_terminated_length": 560.5, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.1072, "grad_norm": 0.11083574593067169, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6000562.0, "reward": 0.8299946784973145, "reward_std": 0.6702513694763184, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5349875092506409, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.7543749809265137, "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": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 554.1875, "completions/mean_terminated_length": 554.1875, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.1074, "grad_norm": 0.06315018236637115, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6013877.0, "reward": 0.5642198324203491, "reward_std": 0.33733800053596497, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19093750417232513, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9056249856948853, "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": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 438.375, "completions/mean_terminated_length": 438.375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1076, "grad_norm": 0.021173710003495216, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6024843.0, "reward": 0.5028887391090393, "reward_std": 0.13567282259464264, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06827499717473984, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 198.0, "completions/max_terminated_length": 198.0, "completions/mean_length": 169.375, "completions/mean_terminated_length": 169.375, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.1078, "grad_norm": 0.1531461775302887, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6031297.0, "reward": 1.3212323188781738, "reward_std": 0.38325849175453186, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8299624919891357, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.971250057220459, "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": 934.0, "completions/mean_length": 612.5, "completions/mean_terminated_length": 585.0667114257812, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.108, "grad_norm": 0.06552388519048691, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6044953.0, "reward": 0.92730712890625, "reward_std": 0.5352684259414673, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47961246967315674, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8899999856948853, "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": 809.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 596.0, "completions/mean_terminated_length": 596.0, "completions/min_length": 450.0, "completions/min_terminated_length": 450.0, "epoch": 0.1082, "grad_norm": 0.051810890436172485, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6057993.0, "reward": 0.5878915786743164, "reward_std": 0.43146607279777527, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2382812350988388, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.921875, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 636.0, "completions/mean_terminated_length": 610.1333618164062, "completions/min_length": 470.0, "completions/min_terminated_length": 470.0, "epoch": 0.1084, "grad_norm": 0.06022738665342331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6072153.0, "reward": 1.0634918212890625, "reward_std": 0.5358243584632874, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.626981258392334, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8818750381469727, "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": 725.0, "completions/max_terminated_length": 725.0, "completions/mean_length": 505.9375, "completions/mean_terminated_length": 505.9375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.1086, "grad_norm": 0.06588944047689438, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6083992.0, "reward": 0.6174229383468628, "reward_std": 0.4035341441631317, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.35984376072883606, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7468750476837158, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 431.875, "completions/mean_terminated_length": 347.2857360839844, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1088, "grad_norm": 0.03173676133155823, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6095062.0, "reward": 0.48209789395332336, "reward_std": 0.1969229131937027, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08919374644756317, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8268749713897705, "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": 0.0, "completions/max_length": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 597.0, "completions/mean_terminated_length": 597.0, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 0.109, "grad_norm": 0.06778270751237869, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6108582.0, "reward": 0.7267976403236389, "reward_std": 0.5075939297676086, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.39109373092651367, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8093750476837158, "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": 597.0, "completions/max_terminated_length": 597.0, "completions/mean_length": 439.25, "completions/mean_terminated_length": 439.25, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.1092, "grad_norm": 0.10456542670726776, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6119242.0, "reward": 1.2848728895187378, "reward_std": 0.47435569763183594, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8197437524795532, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.965624988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 651.375, "completions/mean_terminated_length": 626.5333862304688, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.1094, "grad_norm": 0.03309286758303642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6133712.0, "reward": 0.5263914465904236, "reward_std": 0.3619634807109833, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17778125405311584, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7950000166893005, "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": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 448.5625, "completions/mean_terminated_length": 448.5625, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.1096, "grad_norm": 0.03722715750336647, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6145673.0, "reward": 0.4698790907859802, "reward_std": 0.1840735226869583, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06475624442100525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 409.9375, "completions/mean_terminated_length": 409.9375, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.1098, "grad_norm": 0.045791707932949066, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6157448.0, "reward": 0.5635823011398315, "reward_std": 0.2839342951774597, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12716248631477356, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9662500619888306, "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": 0.0, "completions/max_length": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 245.6875, "completions/mean_terminated_length": 245.6875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.11, "grad_norm": 0.04817677289247513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6165011.0, "reward": 1.4399168491363525, "reward_std": 0.23764507472515106, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9423312544822693, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706249833106995, "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": 596.0, "completions/max_terminated_length": 596.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.1102, "grad_norm": 0.09491610527038574, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6174856.0, "reward": 0.8689602613449097, "reward_std": 0.5214349627494812, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42541876435279846, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 596.875, "completions/mean_terminated_length": 568.4000244140625, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.1104, "grad_norm": 0.08868972212076187, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6188374.0, "reward": 0.7637104392051697, "reward_std": 0.5096266865730286, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.40241873264312744, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7737500071525574, "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": 0.0, "completions/max_length": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 282.8125, "completions/mean_terminated_length": 282.8125, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 0.1106, "grad_norm": 0.1387639045715332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6196707.0, "reward": 0.974429190158844, "reward_std": 0.5135948657989502, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5738562345504761, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8100000619888306, "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": 0.0, "completions/max_length": 1007.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 624.75, "completions/mean_terminated_length": 624.75, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.1108, "grad_norm": 0.031242728233337402, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6210639.0, "reward": 0.70218825340271, "reward_std": 0.2424565851688385, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46687501668930054, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.675000011920929, "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": 0.0, "completions/max_length": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 441.5, "completions/mean_terminated_length": 441.5, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.111, "grad_norm": 0.058539215475320816, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6222711.0, "reward": 0.7646102905273438, "reward_std": 0.3834952414035797, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.34171873331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 619.1875, "completions/mean_terminated_length": 592.2000122070312, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1112, "grad_norm": 0.0792865976691246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6236810.0, "reward": 0.860938549041748, "reward_std": 0.4910637438297272, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.47187501192092896, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8187500238418579, "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": 663.0, "completions/max_terminated_length": 663.0, "completions/mean_length": 446.375, "completions/mean_terminated_length": 446.375, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.1114, "grad_norm": 0.03534979373216629, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6248640.0, "reward": 0.5400012135505676, "reward_std": 0.18702256679534912, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14249998331069946, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.893750011920929, "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": 380.0, "completions/max_terminated_length": 380.0, "completions/mean_length": 305.5625, "completions/mean_terminated_length": 305.5625, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.1116, "grad_norm": 0.09096165746450424, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6258153.0, "reward": 0.6619506478309631, "reward_std": 0.38320648670196533, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5113999843597412, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.41187500953674316, "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": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 408.625, "completions/mean_terminated_length": 408.625, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.1118, "grad_norm": 0.040031448006629944, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6268979.0, "reward": 0.4527360200881958, "reward_std": 0.17816436290740967, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03046874888241291, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9812500476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 468.0625, "completions/mean_terminated_length": 431.0000305175781, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.112, "grad_norm": 0.06984955072402954, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6281156.0, "reward": 0.6353076696395874, "reward_std": 0.4627387225627899, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.20811250805854797, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8612500429153442, "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": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 455.8125, "completions/mean_terminated_length": 455.8125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1122, "grad_norm": 0.11041788011789322, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6292433.0, "reward": 0.9771884679794312, "reward_std": 0.6372501254081726, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5793750286102295, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 379.3125, "completions/mean_terminated_length": 379.3125, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.1124, "grad_norm": 0.08420172333717346, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6302182.0, "reward": 1.318894863128662, "reward_std": 0.3869360685348511, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8252875208854675, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 425.5, "completions/mean_terminated_length": 425.5, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.1126, "grad_norm": 0.08265632390975952, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6313230.0, "reward": 0.7195318341255188, "reward_std": 0.4022165834903717, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5640624761581421, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4625000059604645, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 401.1875, "completions/mean_terminated_length": 359.66668701171875, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.1128, "grad_norm": 0.1188451275229454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6323521.0, "reward": 0.6698322296142578, "reward_std": 0.4554540812969208, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27716249227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8768749833106995, "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": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 402.3125, "completions/mean_terminated_length": 402.3125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.113, "grad_norm": 0.0914023146033287, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6334086.0, "reward": 0.7146103978157043, "reward_std": 0.5016293525695801, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30421876907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.953125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 332.75, "completions/mean_terminated_length": 286.66668701171875, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1132, "grad_norm": 0.1306586116552353, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6343346.0, "reward": 0.9683228731155396, "reward_std": 0.5806036591529846, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5616437792778015, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8868750333786011, "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": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 283.125, "completions/mean_terminated_length": 283.125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.1134, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6351940.0, "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": 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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 303.5625, "completions/mean_terminated_length": 303.5625, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.1136, "grad_norm": 0.09529805928468704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6360525.0, "reward": 1.187023401260376, "reward_std": 0.4425487220287323, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7490437626838684, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8731249570846558, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 474.8125, "completions/mean_terminated_length": 474.8125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.1138, "grad_norm": 0.06585410237312317, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6371738.0, "reward": 1.0191792249679565, "reward_std": 0.496052622795105, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5383563041687012, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 333.3125, "completions/mean_terminated_length": 333.3125, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.114, "grad_norm": 0.10195768624544144, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6380879.0, "reward": 0.9867168664932251, "reward_std": 0.5455206632614136, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5359312295913696, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.96812504529953, "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": 0.0, "completions/max_length": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 560.5, "completions/mean_terminated_length": 560.5, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 0.1142, "grad_norm": 0.05889173224568367, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6394119.0, "reward": 0.6845916509628296, "reward_std": 0.33729052543640137, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24418124556541443, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9131250381469727, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 625.0, "completions/mean_length": 438.5625, "completions/mean_terminated_length": 399.5333557128906, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.1144, "grad_norm": 0.07078737020492554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6405296.0, "reward": 0.6582821607589722, "reward_std": 0.45590940117836, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2540625035762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8925000429153442, "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": 407.0, "completions/max_terminated_length": 407.0, "completions/mean_length": 286.0625, "completions/mean_terminated_length": 286.0625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1146, "grad_norm": 0.11202732473611832, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6413665.0, "reward": 0.648750901222229, "reward_std": 0.4588828980922699, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23499999940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 550.625, "completions/mean_terminated_length": 519.0667114257812, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.1148, "grad_norm": 0.07714279741048813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6427003.0, "reward": 0.9298446178436279, "reward_std": 0.5346524119377136, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4846875071525574, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "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": 931.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 625.5625, "completions/mean_terminated_length": 625.5625, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.115, "grad_norm": 0.038783226162195206, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6440756.0, "reward": 0.5309354066848755, "reward_std": 0.24575714766979218, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18686875700950623, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8556250333786011, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 562.6875, "completions/mean_terminated_length": 408.91668701171875, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 0.1152, "grad_norm": 0.05621479079127312, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6453327.0, "reward": 0.8126565217971802, "reward_std": 0.48910239338874817, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.3128125071525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.07500000298023224, "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": 386.0, "completions/max_terminated_length": 386.0, "completions/mean_length": 314.375, "completions/mean_terminated_length": 314.375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1154, "grad_norm": 0.11407746374607086, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6462325.0, "reward": 0.580641508102417, "reward_std": 0.509896993637085, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2237812578678131, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 250.875, "completions/mean_terminated_length": 250.875, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 0.1156, "grad_norm": 0.006732560228556395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6470355.0, "reward": 0.5378011465072632, "reward_std": 0.020081663504242897, "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": 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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 277.4375, "completions/mean_terminated_length": 277.4375, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1158, "grad_norm": 0.09382927417755127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6478666.0, "reward": 0.6124985218048096, "reward_std": 0.37113717198371887, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1624937355518341, "rewards/confidence_one_or_zero": 0.625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9806250333786011, "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": 0.0, "completions/max_length": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 294.5, "completions/mean_terminated_length": 294.5, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.116, "grad_norm": 0.13246795535087585, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6487058.0, "reward": 1.3459229469299316, "reward_std": 0.4301203489303589, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8793437480926514, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 0.0, "completions/max_length": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 409.0625, "completions/mean_terminated_length": 409.0625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1162, "grad_norm": 0.06520278751850128, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6497619.0, "reward": 0.42265763878822327, "reward_std": 0.2668575048446655, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09531249850988388, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.940625011920929, "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": 702.0, "completions/mean_length": 560.8125, "completions/mean_terminated_length": 529.933349609375, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.1164, "grad_norm": 0.07749629020690918, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6510768.0, "reward": 0.991538405418396, "reward_std": 0.6247778534889221, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6080750226974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8600000143051147, "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": 542.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 344.75, "completions/mean_terminated_length": 344.75, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1166, "grad_norm": 0.07933183759450912, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6519916.0, "reward": 0.6866356134414673, "reward_std": 0.42710766196250916, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2482687532901764, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 0.0, "completions/max_length": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 435.1875, "completions/mean_terminated_length": 435.1875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.1168, "grad_norm": 0.14696671068668365, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6530847.0, "reward": 0.7560823559761047, "reward_std": 0.4642712473869324, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3246625065803528, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 0.0, "completions/max_length": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 481.4375, "completions/mean_terminated_length": 481.4375, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.117, "grad_norm": 0.07895199954509735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6543270.0, "reward": 0.81546950340271, "reward_std": 0.40288347005844116, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.25593751668930054, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5874999761581421, "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": 0.0, "completions/max_length": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 284.0, "completions/mean_terminated_length": 284.0, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.1172, "grad_norm": 0.12847857177257538, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6552518.0, "reward": 1.0779259204864502, "reward_std": 0.4935741722583771, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5933500528335571, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 247.0, "completions/max_terminated_length": 247.0, "completions/mean_length": 170.25, "completions/mean_terminated_length": 170.25, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.1174, "grad_norm": 0.06985277682542801, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6558746.0, "reward": 1.438010573387146, "reward_std": 0.24485914409160614, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.938518762588501, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731249809265137, "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": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 376.3125, "completions/mean_terminated_length": 376.3125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1176, "grad_norm": 0.07309463620185852, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6568735.0, "reward": 0.5958603620529175, "reward_std": 0.3978375494480133, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19171875715255737, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.949999988079071, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 672.1875, "completions/mean_terminated_length": 621.9285888671875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.1178, "grad_norm": 0.03920311853289604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6584162.0, "reward": 0.7151569128036499, "reward_std": 0.3898383378982544, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.49281251430511475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5062499642372131, "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": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 336.875, "completions/mean_terminated_length": 336.875, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.118, "grad_norm": 0.03812841698527336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6594096.0, "reward": 0.48164182901382446, "reward_std": 0.19428898394107819, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08828124403953552, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.949999988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 643.0, "completions/mean_length": 378.375, "completions/mean_terminated_length": 335.3333435058594, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1182, "grad_norm": 0.10838409513235092, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6604022.0, "reward": 0.8134133815765381, "reward_std": 0.4969937801361084, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3768250048160553, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.89124995470047, "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": 492.0, "completions/max_terminated_length": 492.0, "completions/mean_length": 372.3125, "completions/mean_terminated_length": 372.3125, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.1184, "grad_norm": 0.10322681814432144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6614651.0, "reward": 1.1033446788787842, "reward_std": 0.5374904870986938, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6441875100135803, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9356249570846558, "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": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 579.6875, "completions/mean_terminated_length": 579.6875, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.1186, "grad_norm": 0.06809963285923004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6627798.0, "reward": 1.1032634973526, "reward_std": 0.5430606603622437, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6440250277519226, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9725000262260437, "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": 302.0, "completions/max_terminated_length": 302.0, "completions/mean_length": 213.625, "completions/mean_terminated_length": 213.625, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.1188, "grad_norm": 0.1621810495853424, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6634912.0, "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": 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": 354.0, "completions/max_terminated_length": 354.0, "completions/mean_length": 268.5, "completions/mean_terminated_length": 268.5, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.119, "grad_norm": 0.15746629238128662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6643368.0, "reward": 1.3440477848052979, "reward_std": 0.4337179660797119, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8755937218666077, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 716.0, "completions/mean_terminated_length": 613.3333740234375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.1192, "grad_norm": 0.03367193788290024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6658744.0, "reward": 0.3342946171760559, "reward_std": 0.2685360312461853, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.04358749836683273, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6681249737739563, "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": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 506.1875, "completions/mean_terminated_length": 506.1875, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.1194, "grad_norm": 0.07120499759912491, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6670667.0, "reward": 0.6443700194358826, "reward_std": 0.4601436257362366, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.22623750567436218, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9706250429153442, "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": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 264.9375, "completions/mean_terminated_length": 264.9375, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.1196, "grad_norm": 0.13261796534061432, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6678378.0, "reward": 0.9597259759902954, "reward_std": 0.49156758189201355, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4819500148296356, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 286.0, "completions/max_terminated_length": 286.0, "completions/mean_length": 217.625, "completions/mean_terminated_length": 217.625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.1198, "grad_norm": 0.007650627288967371, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6686548.0, "reward": 0.5311853885650635, "reward_std": 0.019159449264407158, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06236875057220459, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 488.0, "completions/max_terminated_length": 488.0, "completions/mean_length": 310.125, "completions/mean_terminated_length": 310.125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.12, "grad_norm": 0.12412841618061066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6695814.0, "reward": 0.8653886318206787, "reward_std": 0.5241255760192871, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41827499866485596, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 425.9375, "completions/mean_terminated_length": 425.9375, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1202, "grad_norm": 0.09133480489253998, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6707029.0, "reward": 0.9616071581840515, "reward_std": 0.5231252312660217, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5482125282287598, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9018750190734863, "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": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 308.4375, "completions/mean_terminated_length": 308.4375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.1204, "grad_norm": 0.23698849976062775, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6716492.0, "reward": 0.8415355682373047, "reward_std": 0.5596482753753662, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43306875228881836, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.953125, "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": 0.0, "completions/max_length": 449.0, "completions/max_terminated_length": 449.0, "completions/mean_length": 329.75, "completions/mean_terminated_length": 329.75, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1206, "grad_norm": 0.08780937641859055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6725336.0, "reward": 1.322091817855835, "reward_std": 0.3810756206512451, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8316812515258789, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 386.0625, "completions/mean_terminated_length": 386.0625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.1208, "grad_norm": 0.09486810863018036, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6735673.0, "reward": 0.8387008905410767, "reward_std": 0.49032384157180786, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.42739999294281006, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9131249785423279, "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": 526.0, "completions/max_terminated_length": 526.0, "completions/mean_length": 405.875, "completions/mean_terminated_length": 405.875, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.121, "grad_norm": 0.06989444047212601, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6746423.0, "reward": 0.7741135358810425, "reward_std": 0.4324597120285034, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2982249855995178, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 0.0, "completions/max_length": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 488.0625, "completions/mean_terminated_length": 488.0625, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.1212, "grad_norm": 0.06629155576229095, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6758216.0, "reward": 1.2398102283477783, "reward_std": 0.37054216861724854, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.8546187877655029, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7831249833106995, "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": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 392.5625, "completions/mean_terminated_length": 392.5625, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.1214, "grad_norm": 0.09019933640956879, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6767889.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 547.75, "completions/mean_terminated_length": 516.0, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.1216, "grad_norm": 0.026776567101478577, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6780413.0, "reward": 0.51343834400177, "reward_std": 0.2207985669374466, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15187500417232513, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8468749523162842, "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": 283.0, "completions/max_terminated_length": 283.0, "completions/mean_length": 220.9375, "completions/mean_terminated_length": 220.9375, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 0.1218, "grad_norm": 0.1255211979150772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6787820.0, "reward": 0.7683634757995605, "reward_std": 0.4364698827266693, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2867249846458435, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "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": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 228.8125, "completions/mean_terminated_length": 228.8125, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.122, "grad_norm": 0.08766336739063263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6795545.0, "reward": 1.3801977634429932, "reward_std": 0.31583961844444275, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8853937387466431, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9368749856948853, "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": 932.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 604.6875, "completions/mean_terminated_length": 604.6875, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.1222, "grad_norm": 0.06435026973485947, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6808740.0, "reward": 0.5020322799682617, "reward_std": 0.34427210688591003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1290625035762787, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8931249976158142, "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": 668.0, "completions/max_terminated_length": 668.0, "completions/mean_length": 474.9375, "completions/mean_terminated_length": 474.9375, "completions/min_length": 344.0, "completions/min_terminated_length": 344.0, "epoch": 0.1224, "grad_norm": 0.07514443248510361, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6820691.0, "reward": 1.1980634927749634, "reward_std": 0.4616613984107971, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7086250185966492, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 815.0, "completions/max_terminated_length": 815.0, "completions/mean_length": 530.0625, "completions/mean_terminated_length": 530.0625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1226, "grad_norm": 0.022225627675652504, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6833364.0, "reward": 0.5499228239059448, "reward_std": 0.16852566599845886, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16234374046325684, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.90625, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 396.875, "completions/mean_terminated_length": 396.875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1228, "grad_norm": 0.1350259631872177, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6844226.0, "reward": 0.8957039713859558, "reward_std": 0.6014772057533264, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5414062738418579, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.903124988079071, "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": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 424.0625, "completions/mean_terminated_length": 424.0625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.123, "grad_norm": 0.08703265339136124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6855299.0, "reward": 0.9027853608131409, "reward_std": 0.4773139953613281, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43056872487068176, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 577.0625, "completions/mean_terminated_length": 577.0625, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.1232, "grad_norm": 0.06572645157575607, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6868820.0, "reward": 1.0187666416168213, "reward_std": 0.5290852189064026, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6000312566757202, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9093749523162842, "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": 0.0, "completions/max_length": 409.0, "completions/max_terminated_length": 409.0, "completions/mean_length": 296.4375, "completions/mean_terminated_length": 296.4375, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.1234, "grad_norm": 0.12162153422832489, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6877339.0, "reward": 1.1629574298858643, "reward_std": 0.5293381214141846, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7009124755859375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9668749570846558, "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": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 370.125, "completions/mean_terminated_length": 370.125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1236, "grad_norm": 0.0976267084479332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6887181.0, "reward": 1.1058447360992432, "reward_std": 0.539380669593811, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6491875648498535, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9543750286102295, "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": 0.0, "completions/max_length": 760.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 438.625, "completions/mean_terminated_length": 438.625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.1238, "grad_norm": 0.11410930752754211, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6898247.0, "reward": 0.7496852874755859, "reward_std": 0.46098199486732483, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31186872720718384, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9337500333786011, "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": 432.0, "completions/max_terminated_length": 432.0, "completions/mean_length": 378.125, "completions/mean_terminated_length": 378.125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.124, "grad_norm": 0.07617902010679245, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6908281.0, "reward": 1.1333262920379639, "reward_std": 0.4885174036026001, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6416500210762024, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "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": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 507.1875, "completions/mean_terminated_length": 507.1875, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.1242, "grad_norm": 0.06781478226184845, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6920940.0, "reward": 0.7345322370529175, "reward_std": 0.4993654489517212, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3440625071525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8718749284744263, "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": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 298.6875, "completions/mean_terminated_length": 298.6875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.1244, "grad_norm": 0.06788187474012375, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6929319.0, "reward": 1.3529198169708252, "reward_std": 0.32483235001564026, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8933374881744385, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9137499928474426, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 587.9375, "completions/mean_terminated_length": 525.6428833007812, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.1246, "grad_norm": 0.02947228215634823, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6943574.0, "reward": 0.5421884059906006, "reward_std": 0.26046279072761536, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20937499403953552, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7062499523162842, "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": 0.0, "completions/max_length": 351.0, "completions/max_terminated_length": 351.0, "completions/mean_length": 277.125, "completions/mean_terminated_length": 277.125, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.1248, "grad_norm": 0.08027306199073792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6951656.0, "reward": 0.6222604513168335, "reward_std": 0.3676440119743347, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18201875686645508, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 375.6875, "completions/mean_terminated_length": 375.6875, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.125, "grad_norm": 0.0669044479727745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 6961795.0, "reward": 1.314469814300537, "reward_std": 0.3975299000740051, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8164374828338623, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9737499952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 447.3125, "completions/mean_terminated_length": 408.86669921875, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1252, "grad_norm": 0.11392996460199356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6973016.0, "reward": 0.6057818531990051, "reward_std": 0.3918880522251129, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.46156251430511475, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5, "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": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 618.8125, "completions/mean_terminated_length": 618.8125, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.1254, "grad_norm": 0.07851919531822205, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 6989269.0, "reward": 0.6206259727478027, "reward_std": 0.5796339511871338, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3037499785423279, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 709.875, "completions/mean_terminated_length": 567.0909423828125, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.1256, "grad_norm": 0.05198463797569275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7004627.0, "reward": 0.36265695095062256, "reward_std": 0.29809391498565674, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1003124937415123, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.6281249523162842, "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": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 311.0, "completions/mean_terminated_length": 311.0, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.1258, "grad_norm": 0.12158756703138351, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7013411.0, "reward": 0.9188200235366821, "reward_std": 0.5443873405456543, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4626374840736389, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9731249809265137, "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": 0.0, "completions/max_length": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 504.625, "completions/mean_terminated_length": 504.625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.126, "grad_norm": 0.06773445755243301, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7025245.0, "reward": 0.5882043838500977, "reward_std": 0.4059191644191742, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23890623450279236, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.846875011920929, "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": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 487.4375, "completions/mean_terminated_length": 487.4375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.1262, "grad_norm": 0.04976171255111694, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7037828.0, "reward": 0.3921012282371521, "reward_std": 0.23464900255203247, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.034199997782707214, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 866.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 637.5625, "completions/mean_terminated_length": 637.5625, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.1264, "grad_norm": 0.06116590276360512, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7052413.0, "reward": 0.6983447074890137, "reward_std": 0.42134320735931396, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.27168750762939453, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9481250047683716, "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": 732.0, "completions/max_terminated_length": 732.0, "completions/mean_length": 482.75, "completions/mean_terminated_length": 482.75, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.1266, "grad_norm": 0.08892222493886948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7063993.0, "reward": 1.1244323253631592, "reward_std": 0.5162705183029175, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6863625049591064, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.925000011920929, "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": 904.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 534.4375, "completions/mean_terminated_length": 534.4375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1268, "grad_norm": 0.08761559426784515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7076560.0, "reward": 0.8895198106765747, "reward_std": 0.5206671953201294, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4665374755859375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8487499952316284, "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": 0.0, "completions/max_length": 841.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 425.0625, "completions/mean_terminated_length": 425.0625, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.127, "grad_norm": 0.010931686498224735, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7087441.0, "reward": 0.5788291692733765, "reward_std": 0.05767424404621124, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1576562523841858, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9156249761581421, "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": 579.0, "completions/max_terminated_length": 579.0, "completions/mean_length": 457.4375, "completions/mean_terminated_length": 457.4375, "completions/min_length": 358.0, "completions/min_terminated_length": 358.0, "epoch": 0.1272, "grad_norm": 0.08475590497255325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7098936.0, "reward": 1.2834135293960571, "reward_std": 0.4763982892036438, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8168250322341919, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9568749666213989, "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": 0.0, "completions/max_length": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 298.875, "completions/mean_terminated_length": 298.875, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.1274, "grad_norm": 0.1825675070285797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7107638.0, "reward": 0.8340541124343872, "reward_std": 0.5614177584648132, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4181062579154968, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9449999928474426, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 755.0, "completions/mean_length": 463.3125, "completions/mean_terminated_length": 425.933349609375, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1276, "grad_norm": 0.09351102262735367, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7118731.0, "reward": 0.7769509553909302, "reward_std": 0.5346866250038147, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36640000343322754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8337500095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 531.875, "completions/mean_terminated_length": 499.0666809082031, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.1278, "grad_norm": 0.07345787435770035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7131689.0, "reward": 0.8650636076927185, "reward_std": 0.524963915348053, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4176250100135803, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9125000238418579, "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": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 363.4375, "completions/mean_terminated_length": 363.4375, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 0.128, "grad_norm": 0.09772954881191254, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7141536.0, "reward": 0.9645947813987732, "reward_std": 0.48698657751083374, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.49168750643730164, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 731.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 470.75, "completions/mean_terminated_length": 470.75, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.1282, "grad_norm": 0.0894789770245552, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7152924.0, "reward": 1.1100791692733765, "reward_std": 0.4945647716522217, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.720156192779541, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8218750357627869, "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": 0.0, "completions/max_length": 265.0, "completions/max_terminated_length": 265.0, "completions/mean_length": 216.875, "completions/mean_terminated_length": 216.875, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.1284, "grad_norm": 0.09604834765195847, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7160426.0, "reward": 1.3411699533462524, "reward_std": 0.35001444816589355, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8698375225067139, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9337500333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 580.4375, "completions/mean_terminated_length": 550.86669921875, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.1286, "grad_norm": 0.1028658077120781, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7173505.0, "reward": 0.79468834400177, "reward_std": 0.5702214241027832, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40187498927116394, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.706250011920929, "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": 589.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 438.5625, "completions/mean_terminated_length": 438.5625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.1288, "grad_norm": 0.07736636698246002, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7184362.0, "reward": 0.8317074775695801, "reward_std": 0.46505850553512573, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35091251134872437, "rewards/confidence_one_or_zero": 0.3125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 460.375, "completions/mean_terminated_length": 422.8000183105469, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.129, "grad_norm": 0.08447112143039703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7196272.0, "reward": 0.8446100950241089, "reward_std": 0.48245900869369507, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5642187595367432, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.699999988079071, "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": 556.0, "completions/max_terminated_length": 556.0, "completions/mean_length": 436.1875, "completions/mean_terminated_length": 436.1875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.1292, "grad_norm": 0.025832761079072952, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7207363.0, "reward": 0.501129150390625, "reward_std": 0.1348438262939453, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06475624442100525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 364.0, "completions/max_terminated_length": 364.0, "completions/mean_length": 248.875, "completions/mean_terminated_length": 248.875, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.1294, "grad_norm": 0.14883941411972046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7215249.0, "reward": 0.8402291536331177, "reward_std": 0.45893433690071106, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3679562509059906, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 306.0625, "completions/mean_terminated_length": 306.0625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.1296, "grad_norm": 0.1104346215724945, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7223954.0, "reward": 1.2583353519439697, "reward_std": 0.43051961064338684, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7666687965393066, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "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": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 611.1875, "completions/mean_terminated_length": 611.1875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.1298, "grad_norm": 0.022093966603279114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7238869.0, "reward": 0.5082042813301086, "reward_std": 0.13762035965919495, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07890624552965164, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 981.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 583.5625, "completions/mean_terminated_length": 583.5625, "completions/min_length": 415.0, "completions/min_terminated_length": 415.0, "epoch": 0.13, "grad_norm": 0.09062144160270691, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7251982.0, "reward": 1.1084322929382324, "reward_std": 0.5367293953895569, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6543624997138977, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 0.0, "completions/max_length": 443.0, "completions/max_terminated_length": 443.0, "completions/mean_length": 309.5625, "completions/mean_terminated_length": 309.5625, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.1302, "grad_norm": 0.05862443521618843, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7260919.0, "reward": 0.5836728811264038, "reward_std": 0.28624337911605835, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16734375059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9393749833106995, "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": 0.0, "completions/max_length": 519.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 400.75, "completions/mean_terminated_length": 400.75, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.1304, "grad_norm": 0.06472985446453094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7271683.0, "reward": 1.3809884786605835, "reward_std": 0.3217903673648834, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8869749903678894, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "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": 738.0, "completions/max_terminated_length": 738.0, "completions/mean_length": 469.75, "completions/mean_terminated_length": 469.75, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.1306, "grad_norm": 0.05460628122091293, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7283391.0, "reward": 0.8342190980911255, "reward_std": 0.27642568945884705, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7309374809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 532.3125, "completions/mean_terminated_length": 499.5333557128906, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1308, "grad_norm": 0.05349404737353325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7295748.0, "reward": 0.5892196893692017, "reward_std": 0.3541732132434845, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2409374862909317, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7906249761581421, "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": 450.0, "completions/max_terminated_length": 450.0, "completions/mean_length": 280.6875, "completions/mean_terminated_length": 280.6875, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 0.131, "grad_norm": 0.09194104373455048, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7304367.0, "reward": 0.7350761890411377, "reward_std": 0.3849906623363495, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2826499938964844, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8899999856948853, "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": 1009.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 518.6875, "completions/mean_terminated_length": 518.6875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.1312, "grad_norm": 0.06288322806358337, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7316602.0, "reward": 0.62406325340271, "reward_std": 0.35577043890953064, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.43562501668930054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6031249761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 508.5625, "completions/mean_terminated_length": 474.20001220703125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.1314, "grad_norm": 0.0835479199886322, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7329043.0, "reward": 0.6135165095329285, "reward_std": 0.4466216564178467, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28953123092651367, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6968749761581421, "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": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 357.625, "completions/mean_terminated_length": 357.625, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.1316, "grad_norm": 0.0940885841846466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7338589.0, "reward": 1.130704641342163, "reward_std": 0.4917774796485901, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6364062428474426, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781249761581421, "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": 733.0, "completions/max_terminated_length": 733.0, "completions/mean_length": 414.625, "completions/mean_terminated_length": 414.625, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.1318, "grad_norm": 0.04001566767692566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7349607.0, "reward": 0.6420320868492126, "reward_std": 0.2651483714580536, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.15906250476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.893750011920929, "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": 0.0, "completions/max_length": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 545.125, "completions/mean_terminated_length": 545.125, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.132, "grad_norm": 0.05821075290441513, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7362217.0, "reward": 0.7759383916854858, "reward_std": 0.3258064091205597, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4268749952316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.762499988079071, "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": 512.0, "completions/max_terminated_length": 512.0, "completions/mean_length": 393.4375, "completions/mean_terminated_length": 393.4375, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 0.1322, "grad_norm": 0.06435630470514297, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7373248.0, "reward": 0.6614073514938354, "reward_std": 0.3272806406021118, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1978124976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 325.5, "completions/mean_terminated_length": 325.5, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.1324, "grad_norm": 0.08438950031995773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7382392.0, "reward": 0.675835371017456, "reward_std": 0.3274390697479248, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22666874527931213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9381250143051147, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 528.9375, "completions/mean_terminated_length": 458.21429443359375, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.1326, "grad_norm": 0.12026827782392502, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7395815.0, "reward": 0.6027230024337769, "reward_std": 0.662465512752533, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.33044373989105225, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.6700000166893005, "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": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 581.625, "completions/mean_terminated_length": 581.625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1328, "grad_norm": 0.0774751827120781, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7409057.0, "reward": 0.6740636825561523, "reward_std": 0.5337772369384766, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.28562498092651367, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 302.4375, "completions/mean_terminated_length": 302.4375, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.133, "grad_norm": 0.08867661654949188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7417784.0, "reward": 1.2593448162078857, "reward_std": 0.42958831787109375, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7686874866485596, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 563.25, "completions/mean_terminated_length": 497.4285888671875, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1332, "grad_norm": 0.05700905621051788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7431164.0, "reward": 0.6290632486343384, "reward_std": 0.39594826102256775, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.258124977350235, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.793749988079071, "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": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 458.5625, "completions/mean_terminated_length": 458.5625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.1334, "grad_norm": 0.10354243218898773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7443461.0, "reward": 0.7975618243217468, "reward_std": 0.5963982343673706, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4076216220855713, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8875125646591187, "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": 434.0, "completions/max_terminated_length": 434.0, "completions/mean_length": 323.5, "completions/mean_terminated_length": 323.5, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1336, "grad_norm": 0.05613435059785843, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7452765.0, "reward": 0.73414146900177, "reward_std": 0.24019935727119446, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.40578123927116394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7406250238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 541.8125, "completions/mean_terminated_length": 509.66668701171875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1338, "grad_norm": 0.10263678431510925, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7465082.0, "reward": 0.8875008225440979, "reward_std": 0.5513102412223816, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5249999761581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7749999761581421, "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": 927.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 694.875, "completions/mean_terminated_length": 694.875, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.134, "grad_norm": 0.07991673797369003, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7480424.0, "reward": 0.6041417121887207, "reward_std": 0.47591379284858704, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3332812190055847, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7900000214576721, "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": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 473.1875, "completions/mean_terminated_length": 473.1875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.1342, "grad_norm": 0.0665983036160469, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7492171.0, "reward": 1.07718825340271, "reward_std": 0.44638463854789734, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.7793750166893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6625000238418579, "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": 278.0, "completions/max_terminated_length": 278.0, "completions/mean_length": 181.625, "completions/mean_terminated_length": 181.625, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "epoch": 0.1344, "grad_norm": 0.012631782330572605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7500293.0, "reward": 0.5285729765892029, "reward_std": 0.02697606198489666, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05714374780654907, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706249833106995, "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": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 391.25, "completions/mean_terminated_length": 391.25, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.1346, "grad_norm": 0.08562875539064407, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7510345.0, "reward": 0.9754667282104492, "reward_std": 0.47930049896240234, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5134312510490417, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "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": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 549.1875, "completions/mean_terminated_length": 549.1875, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.1348, "grad_norm": 0.0960456132888794, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7523740.0, "reward": 1.1044197082519531, "reward_std": 0.5417179465293884, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6463375091552734, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.971875011920929, "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": 372.0, "completions/max_terminated_length": 372.0, "completions/mean_length": 281.25, "completions/mean_terminated_length": 281.25, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.135, "grad_norm": 0.10639510303735733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7532704.0, "reward": 1.0165448188781738, "reward_std": 0.49837765097618103, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5330874919891357, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 816.0, "completions/max_terminated_length": 816.0, "completions/mean_length": 570.125, "completions/mean_terminated_length": 570.125, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.1352, "grad_norm": 0.06057598814368248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7545378.0, "reward": 0.6564853191375732, "reward_std": 0.3677859306335449, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.25046876072883606, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.859375, "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": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 400.5625, "completions/mean_terminated_length": 400.5625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1354, "grad_norm": 0.03208513930439949, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7555403.0, "reward": 0.5479698181152344, "reward_std": 0.1865251362323761, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15843749046325684, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8843749761581421, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 578.5, "completions/mean_terminated_length": 514.857177734375, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.1356, "grad_norm": 0.0583377368748188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7568211.0, "reward": 0.6572538614273071, "reward_std": 0.4113675057888031, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.31450626254081726, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7331249713897705, "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": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 678.75, "completions/mean_terminated_length": 678.75, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.1358, "grad_norm": 0.015592724084854126, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7582895.0, "reward": 0.5202353000640869, "reward_std": 0.1396171897649765, "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": 0.9437500238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 731.0, "completions/mean_length": 519.9375, "completions/mean_terminated_length": 486.3333435058594, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.136, "grad_norm": 0.10887525975704193, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7595438.0, "reward": 0.8910822868347168, "reward_std": 0.5803833603858948, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46966248750686646, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9018750190734863, "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": 551.0, "completions/max_terminated_length": 551.0, "completions/mean_length": 306.8125, "completions/mean_terminated_length": 306.8125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1362, "grad_norm": 0.014059232547879219, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7605179.0, "reward": 0.559826135635376, "reward_std": 0.07067088037729263, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1196499913930893, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9350000619888306, "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": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 520.3125, "completions/mean_terminated_length": 520.3125, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.1364, "grad_norm": 0.05634293332695961, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7617328.0, "reward": 0.6614041328430176, "reward_std": 0.37211552262306213, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2603062391281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 590.0, "completions/max_terminated_length": 590.0, "completions/mean_length": 422.4375, "completions/mean_terminated_length": 422.4375, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.1366, "grad_norm": 0.09165085852146149, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7628311.0, "reward": 1.0832571983337402, "reward_std": 0.4868498146533966, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6040124893188477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 677.0, "completions/mean_terminated_length": 677.0, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.1368, "grad_norm": 0.0455692857503891, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7643911.0, "reward": 0.8128912448883057, "reward_std": 0.3053305149078369, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.500781238079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.684374988079071, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 580.75, "completions/mean_terminated_length": 517.4285888671875, "completions/min_length": 340.0, "completions/min_terminated_length": 340.0, "epoch": 0.137, "grad_norm": 0.10212546586990356, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7657539.0, "reward": 0.7139068841934204, "reward_std": 0.5638213157653809, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5528124570846558, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.4937500059604645, "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": 0.0, "completions/max_length": 621.0, "completions/max_terminated_length": 621.0, "completions/mean_length": 393.3125, "completions/mean_terminated_length": 393.3125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.1372, "grad_norm": 0.049803536385297775, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7667880.0, "reward": 0.573672890663147, "reward_std": 0.2826310098171234, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14734375476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "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": 0.0, "completions/max_length": 308.0, "completions/max_terminated_length": 308.0, "completions/mean_length": 251.125, "completions/mean_terminated_length": 251.125, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.1374, "grad_norm": 0.05414620414376259, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7675674.0, "reward": 0.5856791734695435, "reward_std": 0.2442813664674759, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10885625332593918, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9731250405311584, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 693.625, "completions/mean_terminated_length": 671.6000366210938, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.1376, "grad_norm": 0.04602856934070587, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7692228.0, "reward": 0.6243757009506226, "reward_std": 0.3628036677837372, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4362500309944153, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.4625000059604645, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 726.375, "completions/mean_terminated_length": 683.857177734375, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.1378, "grad_norm": 0.06453432142734528, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7710218.0, "reward": 0.40695416927337646, "reward_std": 0.4498263895511627, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1889062523841858, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.721875011920929, "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": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 488.625, "completions/mean_terminated_length": 488.625, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.138, "grad_norm": 0.0982460081577301, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7722436.0, "reward": 0.8834106922149658, "reward_std": 0.5858104228973389, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45431873202323914, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9831250309944153, "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": 793.0, "completions/mean_length": 542.3125, "completions/mean_terminated_length": 510.20001220703125, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.1382, "grad_norm": 0.04803965613245964, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7735385.0, "reward": 0.6259227991104126, "reward_std": 0.366566002368927, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18934375047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981250524520874, "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": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 674.125, "completions/mean_terminated_length": 674.125, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.1384, "grad_norm": 0.024929547682404518, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7750539.0, "reward": 0.5423446893692017, "reward_std": 0.1825026422739029, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1471875011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8374999761581421, "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": 0.0, "completions/max_length": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 566.1875, "completions/mean_terminated_length": 566.1875, "completions/min_length": 440.0, "completions/min_terminated_length": 440.0, "epoch": 0.1386, "grad_norm": 0.06340152770280838, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7763406.0, "reward": 0.8525761365890503, "reward_std": 0.45368826389312744, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3926500082015991, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "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": 564.0, "completions/max_terminated_length": 564.0, "completions/mean_length": 326.5, "completions/mean_terminated_length": 326.5, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.1388, "grad_norm": 0.132259339094162, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7772358.0, "reward": 0.8784202933311462, "reward_std": 0.4968744218349457, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38183748722076416, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 0.0, "completions/max_length": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 474.25, "completions/mean_terminated_length": 474.25, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.139, "grad_norm": 0.08256487548351288, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7784202.0, "reward": 0.6787135601043701, "reward_std": 0.5324609279632568, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2949249744415283, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 717.5625, "completions/mean_terminated_length": 673.7857666015625, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 0.1392, "grad_norm": 0.04469415917992592, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7799747.0, "reward": 0.48859459161758423, "reward_std": 0.3370116055011749, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.22718751430511475, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6343749761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 355.5625, "completions/mean_terminated_length": 311.0000305175781, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.1394, "grad_norm": 0.1379159837961197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7809468.0, "reward": 1.0694103240966797, "reward_std": 0.5946439504623413, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6388187408447266, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9075000286102295, "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": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 388.8125, "completions/mean_terminated_length": 388.8125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.1396, "grad_norm": 0.09081436693668365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7819753.0, "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": 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": 541.0, "completions/max_terminated_length": 541.0, "completions/mean_length": 455.0625, "completions/mean_terminated_length": 455.0625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.1398, "grad_norm": 0.1286923587322235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7831002.0, "reward": 1.3461384773254395, "reward_std": 0.4302023649215698, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8797750473022461, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.0, "completions/max_length": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 284.0625, "completions/mean_terminated_length": 284.0625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.14, "grad_norm": 0.1353343427181244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7839867.0, "reward": 0.7560133934020996, "reward_std": 0.5363072752952576, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.44952499866485596, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7637500166893005, "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": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 381.8125, "completions/mean_terminated_length": 381.8125, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.1402, "grad_norm": 0.09986304491758347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7849944.0, "reward": 1.2020853757858276, "reward_std": 0.4549238979816437, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7166687250137329, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "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": 645.0, "completions/max_terminated_length": 645.0, "completions/mean_length": 458.0, "completions/mean_terminated_length": 458.0, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.1404, "grad_norm": 0.04636947438120842, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7862024.0, "reward": 0.8498414158821106, "reward_std": 0.28238168358802795, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.637181282043457, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4806250333786011, "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": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 295.1875, "completions/mean_terminated_length": 295.1875, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.1406, "grad_norm": 0.13297919929027557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7870459.0, "reward": 1.1628637313842773, "reward_std": 0.5304946303367615, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7007249593734741, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 0.0, "completions/max_length": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 348.875, "completions/mean_terminated_length": 348.875, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.1408, "grad_norm": 0.10216046869754791, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7879673.0, "reward": 0.8543634414672852, "reward_std": 0.4491865932941437, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3962250053882599, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "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": 775.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 465.5, "completions/mean_terminated_length": 465.5, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.141, "grad_norm": 0.05760176479816437, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7891153.0, "reward": 0.8324223160743713, "reward_std": 0.27429381012916565, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7273437976837158, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.328125, "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": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 379.8125, "completions/mean_terminated_length": 379.8125, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.1412, "grad_norm": 0.005694442894309759, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7900814.0, "reward": 0.5598446726799011, "reward_std": 0.03405528515577316, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.11968749761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 650.1875, "completions/mean_terminated_length": 625.2667236328125, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.1414, "grad_norm": 0.03787446767091751, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7914753.0, "reward": 0.48882919549942017, "reward_std": 0.27444949746131897, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16515624523162842, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.824999988079071, "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": 332.0, "completions/max_terminated_length": 332.0, "completions/mean_length": 256.75, "completions/mean_terminated_length": 256.75, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1416, "grad_norm": 0.16908332705497742, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7922781.0, "reward": 0.8194040656089783, "reward_std": 0.5688174962997437, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.38880616426467896, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9036874771118164, "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": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 313.3125, "completions/mean_terminated_length": 313.3125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1418, "grad_norm": 0.15071731805801392, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7931986.0, "reward": 0.7458603382110596, "reward_std": 0.5029410719871521, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.36671876907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.890625, "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": 0.0, "completions/max_length": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 329.3125, "completions/mean_terminated_length": 329.3125, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.142, "grad_norm": 0.044570598751306534, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7941111.0, "reward": 0.7213257551193237, "reward_std": 0.21841420233249664, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3801499903202057, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7587499618530273, "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": 685.0, "completions/max_terminated_length": 685.0, "completions/mean_length": 377.375, "completions/mean_terminated_length": 377.375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1422, "grad_norm": 0.12136916071176529, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7951277.0, "reward": 0.9784762859344482, "reward_std": 0.5531682372093201, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5194500088691711, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.984375, "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": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 263.25, "completions/mean_terminated_length": 263.25, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.1424, "grad_norm": 0.16021710634231567, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7958993.0, "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": 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": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 357.25, "completions/mean_terminated_length": 357.25, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1426, "grad_norm": 0.07523466646671295, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7969445.0, "reward": 0.4451322555541992, "reward_std": 0.22142033278942108, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07776249945163727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 392.0, "completions/max_terminated_length": 392.0, "completions/mean_length": 282.4375, "completions/mean_terminated_length": 282.4375, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.1428, "grad_norm": 0.13059964776039124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 7978060.0, "reward": 0.902800977230072, "reward_std": 0.47733479738235474, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43059998750686646, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 440.0, "completions/max_terminated_length": 440.0, "completions/mean_length": 323.75, "completions/mean_terminated_length": 323.75, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.143, "grad_norm": 0.13539212942123413, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 7988856.0, "reward": 1.0963424444198608, "reward_std": 0.48382726311683655, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6301819086074829, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9483312368392944, "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": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 472.3125, "completions/mean_terminated_length": 472.3125, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.1432, "grad_norm": 0.09194587171077728, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8003917.0, "reward": 0.7080696821212769, "reward_std": 0.5042005181312561, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2911374866962433, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9625000357627869, "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": 0.0, "completions/max_length": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 407.25, "completions/mean_terminated_length": 407.25, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1434, "grad_norm": 0.11869918555021286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8014753.0, "reward": 0.8856260776519775, "reward_std": 0.4981653690338135, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.45875000953674316, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.859375, "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": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 306.25, "completions/mean_terminated_length": 306.25, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.1436, "grad_norm": 0.002672871109098196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8023349.0, "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": 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": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 313.6875, "completions/mean_terminated_length": 313.6875, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1438, "grad_norm": 0.12159170210361481, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8033056.0, "reward": 0.962722897529602, "reward_std": 0.4886123239994049, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48794373869895935, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 294.0, "completions/max_terminated_length": 294.0, "completions/mean_length": 210.375, "completions/mean_terminated_length": 210.375, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.144, "grad_norm": 0.14598913490772247, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8039974.0, "reward": 0.6218323111534119, "reward_std": 0.3680679202079773, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1811625063419342, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 475.25, "completions/mean_terminated_length": 475.25, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 0.1442, "grad_norm": 0.055234432220458984, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8051130.0, "reward": 0.686082124710083, "reward_std": 0.32767045497894287, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24716249108314514, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9237499833106995, "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": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 338.875, "completions/mean_terminated_length": 338.875, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.1444, "grad_norm": 0.10042949765920639, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8060520.0, "reward": 1.318638563156128, "reward_std": 0.3884744644165039, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8247750401496887, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 337.0, "completions/max_terminated_length": 337.0, "completions/mean_length": 257.375, "completions/mean_terminated_length": 257.375, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1446, "grad_norm": 0.15735796093940735, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8068798.0, "reward": 0.9446228742599487, "reward_std": 0.5962291359901428, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5142437219619751, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 570.375, "completions/mean_terminated_length": 570.375, "completions/min_length": 418.0, "completions/min_terminated_length": 418.0, "epoch": 0.1448, "grad_norm": 0.0452500656247139, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8081732.0, "reward": 0.7106887102127075, "reward_std": 0.39213135838508606, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23387494683265686, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9749374985694885, "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": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 454.6875, "completions/mean_terminated_length": 454.6875, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.145, "grad_norm": 0.10602317750453949, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8094367.0, "reward": 0.9838290214538574, "reward_std": 0.5704257488250732, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5926562547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.856249988079071, "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": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 391.125, "completions/mean_terminated_length": 391.125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.1452, "grad_norm": 0.10900276899337769, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8104641.0, "reward": 1.1421759128570557, "reward_std": 0.4393138289451599, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7218500375747681, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8362500071525574, "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": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 230.8125, "completions/mean_terminated_length": 230.8125, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.1454, "grad_norm": 0.05790659040212631, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8112302.0, "reward": 0.5364572405815125, "reward_std": 0.1843966841697693, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13541249930858612, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8375000357627869, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 622.1875, "completions/mean_terminated_length": 595.4000244140625, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.1456, "grad_norm": 0.040423739701509476, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8126785.0, "reward": 0.8192191123962402, "reward_std": 0.2943230867385864, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7009375095367432, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3187499940395355, "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": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 424.5625, "completions/mean_terminated_length": 424.5625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1458, "grad_norm": 0.06085380166769028, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8137370.0, "reward": 0.7153666019439697, "reward_std": 0.3897916376590729, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24323125183582306, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 328.625, "completions/mean_terminated_length": 328.625, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 0.146, "grad_norm": 0.12682580947875977, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8146500.0, "reward": 1.081466555595398, "reward_std": 0.48903483152389526, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6004312634468079, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "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": 757.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 492.6875, "completions/mean_terminated_length": 492.6875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.1462, "grad_norm": 0.047560546547174454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8158719.0, "reward": 0.5476452708244324, "reward_std": 0.285593718290329, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.0952874943614006, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9793750047683716, "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": 575.0, "completions/max_terminated_length": 575.0, "completions/mean_length": 366.125, "completions/mean_terminated_length": 366.125, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1464, "grad_norm": 0.0455094538629055, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8168737.0, "reward": 0.5385759472846985, "reward_std": 0.18341124057769775, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13964998722076416, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500023841858, "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": 516.0, "completions/max_terminated_length": 516.0, "completions/mean_length": 316.3125, "completions/mean_terminated_length": 316.3125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1466, "grad_norm": 0.11311883479356766, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8177654.0, "reward": 1.048985481262207, "reward_std": 0.5431162118911743, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5979687571525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593749642372131, "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": 705.0, "completions/max_terminated_length": 705.0, "completions/mean_length": 566.0, "completions/mean_terminated_length": 566.0, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.1468, "grad_norm": 0.03357553854584694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8190774.0, "reward": 0.9910160303115845, "reward_std": 0.24698355793952942, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.7945312261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.453125, "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": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 506.0625, "completions/mean_terminated_length": 506.0625, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.147, "grad_norm": 0.051044803112745285, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8203127.0, "reward": 0.7390632629394531, "reward_std": 0.2908697724342346, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3531249761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 611.6875, "completions/mean_terminated_length": 584.2000122070312, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 0.1472, "grad_norm": 0.021581783890724182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8217202.0, "reward": 0.5512509346008301, "reward_std": 0.1883116215467453, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16500000655651093, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "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": 0.0, "completions/max_length": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 406.75, "completions/mean_terminated_length": 406.75, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.1474, "grad_norm": 0.06286878883838654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8227950.0, "reward": 0.5984134674072266, "reward_std": 0.3021553158760071, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19682499766349792, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 292.8125, "completions/mean_terminated_length": 292.8125, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.1476, "grad_norm": 0.13362911343574524, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8236507.0, "reward": 0.5993728041648865, "reward_std": 0.3974834382534027, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1987437605857849, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9462500214576721, "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": 269.0, "completions/max_terminated_length": 269.0, "completions/mean_length": 214.0625, "completions/mean_terminated_length": 214.0625, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.1478, "grad_norm": 0.1187916100025177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8243692.0, "reward": 0.7160510420799255, "reward_std": 0.3888743221759796, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24459999799728394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 741.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 534.375, "completions/mean_terminated_length": 534.375, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.148, "grad_norm": 0.024466371163725853, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8257058.0, "reward": 0.5142979025840759, "reward_std": 0.13823890686035156, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0910937488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 835.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 658.4375, "completions/mean_terminated_length": 658.4375, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.1482, "grad_norm": 0.050424665212631226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8271785.0, "reward": 1.043335199356079, "reward_std": 0.3706260621547699, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.7116687893867493, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5756250023841858, "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": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 284.25, "completions/mean_terminated_length": 284.25, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.1484, "grad_norm": 0.13740231096744537, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8280429.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 428.8125, "completions/mean_terminated_length": 389.13336181640625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1486, "grad_norm": 0.10598690807819366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8291274.0, "reward": 1.0460666418075562, "reward_std": 0.54635089635849, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5921312570571899, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9031250476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 516.625, "completions/mean_terminated_length": 482.8000183105469, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1488, "grad_norm": 0.06880109012126923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8303428.0, "reward": 0.7450756430625916, "reward_std": 0.33226126432418823, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6151500344276428, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.4337500035762787, "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": 0.0, "completions/max_length": 580.0, "completions/max_terminated_length": 580.0, "completions/mean_length": 412.1875, "completions/mean_terminated_length": 412.1875, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.149, "grad_norm": 0.0914190411567688, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8314007.0, "reward": 0.964375913143158, "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": 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": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 321.9375, "completions/mean_terminated_length": 321.9375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.1492, "grad_norm": 0.09398289769887924, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8323174.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 587.6875, "completions/mean_terminated_length": 558.6000366210938, "completions/min_length": 396.0, "completions/min_terminated_length": 396.0, "epoch": 0.1494, "grad_norm": 0.05642089992761612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8338145.0, "reward": 0.7064790725708008, "reward_std": 0.4704505503177643, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.35045623779296875, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7862499952316284, "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": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 435.25, "completions/mean_terminated_length": 435.25, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1496, "grad_norm": 0.055638208985328674, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8348917.0, "reward": 0.7446883916854858, "reward_std": 0.25451961159706116, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4268750250339508, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7374999523162842, "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": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 390.9375, "completions/mean_terminated_length": 390.9375, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1498, "grad_norm": 0.06681331247091293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8358836.0, "reward": 0.7532821893692017, "reward_std": 0.3752110004425049, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3190625011920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.918749988079071, "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": 347.0, "completions/max_terminated_length": 347.0, "completions/mean_length": 246.75, "completions/mean_terminated_length": 246.75, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.15, "grad_norm": 0.15725654363632202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8367216.0, "reward": 1.3462133407592773, "reward_std": 0.4302307069301605, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.879925012588501, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 323.0, "completions/max_terminated_length": 323.0, "completions/mean_length": 264.0625, "completions/mean_terminated_length": 264.0625, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.1502, "grad_norm": 0.09050103276968002, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8375153.0, "reward": 1.3206915855407715, "reward_std": 0.3829898536205292, "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": 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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 657.0625, "completions/mean_terminated_length": 572.3846435546875, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.1504, "grad_norm": 0.03397637978196144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8390066.0, "reward": 0.5393756628036499, "reward_std": 0.29421737790107727, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26625001430511475, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6499999761581421, "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": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 612.375, "completions/mean_terminated_length": 612.375, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.1506, "grad_norm": 0.007057786453515291, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8404520.0, "reward": 0.5511729717254639, "reward_std": 0.03151516616344452, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10234375298023224, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 358.5, "completions/mean_terminated_length": 358.5, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.1508, "grad_norm": 0.1228656992316246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8415088.0, "reward": 0.919826090335846, "reward_std": 0.5431302189826965, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4646500051021576, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 0.0, "completions/max_length": 482.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 358.5, "completions/mean_terminated_length": 358.5, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.151, "grad_norm": 0.10515975952148438, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8424568.0, "reward": 0.9393759965896606, "reward_std": 0.4615452289581299, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5037499666213989, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8937499523162842, "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": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 535.3125, "completions/mean_terminated_length": 535.3125, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1512, "grad_norm": 0.04722274839878082, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8437757.0, "reward": 0.538266658782959, "reward_std": 0.25734174251556396, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20153124630451202, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8506250381469727, "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": 769.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 583.8125, "completions/mean_terminated_length": 583.8125, "completions/min_length": 471.0, "completions/min_terminated_length": 471.0, "epoch": 0.1514, "grad_norm": 0.0904010683298111, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8451210.0, "reward": 1.2489478588104248, "reward_std": 0.44662922620773315, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8728936910629272, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8012499809265137, "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": 333.0, "completions/max_terminated_length": 333.0, "completions/mean_length": 223.5, "completions/mean_terminated_length": 223.5, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.1516, "grad_norm": 0.09038658440113068, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8458370.0, "reward": 1.3758387565612793, "reward_std": 0.33800843358039856, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.876675009727478, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 322.3125, "completions/mean_terminated_length": 322.3125, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.1518, "grad_norm": 0.13825567066669464, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8467319.0, "reward": 0.8708353042602539, "reward_std": 0.520275890827179, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4291687607765198, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9556249976158142, "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": 918.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 604.4375, "completions/mean_terminated_length": 604.4375, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.152, "grad_norm": 0.061977412551641464, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8481070.0, "reward": 0.738813042640686, "reward_std": 0.27045682072639465, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5401249527931213, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5436874628067017, "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": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 332.5625, "completions/mean_terminated_length": 332.5625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1522, "grad_norm": 0.11240732669830322, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8490343.0, "reward": 1.3457822799682617, "reward_std": 0.430066853761673, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8790625333786011, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.953125, "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": 0.0, "completions/max_length": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 423.9375, "completions/mean_terminated_length": 423.9375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.1524, "grad_norm": 0.08631135523319244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8501014.0, "reward": 1.1187291145324707, "reward_std": 0.45827171206474304, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6749563217163086, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.909375011920929, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 721.25, "completions/mean_terminated_length": 701.0667114257812, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.1526, "grad_norm": 0.02980796992778778, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8517386.0, "reward": 0.7938286066055298, "reward_std": 0.2864961624145508, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6501562595367432, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.37812501192092896, "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": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 401.375, "completions/mean_terminated_length": 401.375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.1528, "grad_norm": 0.054918356239795685, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8527664.0, "reward": 0.5767196416854858, "reward_std": 0.28204604983329773, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15343749523162842, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 439.8125, "completions/mean_terminated_length": 439.8125, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.153, "grad_norm": 0.11228638887405396, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8538509.0, "reward": 0.7740635871887207, "reward_std": 0.4695567786693573, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3606249988079071, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 506.625, "completions/mean_terminated_length": 506.625, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.1532, "grad_norm": 0.08657284826040268, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8550679.0, "reward": 1.1352638006210327, "reward_std": 0.4862706959247589, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6455249786376953, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "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": 566.0, "completions/max_terminated_length": 566.0, "completions/mean_length": 414.6875, "completions/mean_terminated_length": 414.6875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1534, "grad_norm": 0.0733584314584732, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8561138.0, "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": 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": 0.0, "completions/max_length": 557.0, "completions/max_terminated_length": 557.0, "completions/mean_length": 391.375, "completions/mean_terminated_length": 391.375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.1536, "grad_norm": 0.06881527602672577, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8571080.0, "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": 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": 460.0, "completions/max_terminated_length": 460.0, "completions/mean_length": 349.1875, "completions/mean_terminated_length": 349.1875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.1538, "grad_norm": 0.09179233759641647, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8581355.0, "reward": 0.8749197721481323, "reward_std": 0.4258140027523041, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.3748375177383423, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8337500095367432, "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": 0.0, "completions/max_length": 423.0, "completions/max_terminated_length": 423.0, "completions/mean_length": 345.8125, "completions/mean_terminated_length": 345.8125, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.154, "grad_norm": 0.11811260133981705, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8590296.0, "reward": 1.11155104637146, "reward_std": 0.5338067412376404, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.660599946975708, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9037500023841858, "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": 0.0, "completions/max_length": 464.0, "completions/max_terminated_length": 464.0, "completions/mean_length": 315.6875, "completions/mean_terminated_length": 315.6875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.1542, "grad_norm": 0.06570662558078766, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8598979.0, "reward": 1.320857286453247, "reward_std": 0.3830723464488983, "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": 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": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 542.8125, "completions/mean_terminated_length": 542.8125, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.1544, "grad_norm": 0.032769326120615005, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8612304.0, "reward": 0.5110949873924255, "reward_std": 0.13900727033615112, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0846875011920929, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 472.0, "completions/mean_terminated_length": 472.0, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.1546, "grad_norm": 0.07952899485826492, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8623696.0, "reward": 0.7177355289459229, "reward_std": 0.38796359300613403, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24796874821186066, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593749642372131, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 789.0, "completions/mean_length": 516.9375, "completions/mean_terminated_length": 483.13336181640625, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.1548, "grad_norm": 0.016024619340896606, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8635999.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 692.8125, "completions/mean_terminated_length": 670.7333984375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.155, "grad_norm": 0.0923897922039032, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8652604.0, "reward": 0.6727350950241089, "reward_std": 0.4954585134983063, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4704687297344208, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5249999761581421, "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": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 346.9375, "completions/mean_terminated_length": 346.9375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 0.1552, "grad_norm": 0.12618179619312286, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8662107.0, "reward": 1.0124887228012085, "reward_std": 0.5031245350837708, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5249750018119812, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9724999666213989, "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": 891.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 454.1875, "completions/mean_terminated_length": 454.1875, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.1554, "grad_norm": 0.0951726883649826, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8673086.0, "reward": 1.0423227548599243, "reward_std": 0.4787522256374359, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5846437215805054, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "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": 545.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 467.125, "completions/mean_terminated_length": 467.125, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.1556, "grad_norm": 0.061082012951374054, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8685104.0, "reward": 0.7634887099266052, "reward_std": 0.4394552409648895, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.27697497606277466, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 637.875, "completions/mean_terminated_length": 637.875, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.1558, "grad_norm": 0.06422919780015945, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8699262.0, "reward": 0.8873445987701416, "reward_std": 0.508352518081665, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4621874988079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.856249988079071, "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": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 397.375, "completions/mean_terminated_length": 397.375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.156, "grad_norm": 0.009056741371750832, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8709284.0, "reward": 0.560457170009613, "reward_std": 0.044485580176115036, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12091249227523804, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9362499713897705, "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": 0.0, "completions/max_length": 788.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 676.4375, "completions/mean_terminated_length": 676.4375, "completions/min_length": 519.0, "completions/min_terminated_length": 519.0, "epoch": 0.1562, "grad_norm": 0.0584864616394043, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8724619.0, "reward": 0.7703131437301636, "reward_std": 0.38609784841537476, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7281249761581421, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.32499998807907104, "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": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 523.0625, "completions/mean_terminated_length": 523.0625, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.1564, "grad_norm": 0.09932796657085419, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8737516.0, "reward": 1.130666732788086, "reward_std": 0.5846500396728516, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.6988312005996704, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.908750057220459, "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": 0.0, "completions/max_length": 938.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 399.1875, "completions/mean_terminated_length": 399.1875, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1566, "grad_norm": 0.10410504788160324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8747583.0, "reward": 1.142991542816162, "reward_std": 0.47539275884628296, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.660981297492981, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "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": 0.0, "completions/max_length": 796.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 605.875, "completions/mean_terminated_length": 605.875, "completions/min_length": 346.0, "completions/min_terminated_length": 346.0, "epoch": 0.1568, "grad_norm": 0.054926756769418716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8762013.0, "reward": 0.6320319175720215, "reward_std": 0.34592482447624207, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4515625238418579, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6156249642372131, "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": 402.0, "completions/max_terminated_length": 402.0, "completions/mean_length": 309.9375, "completions/mean_terminated_length": 309.9375, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.157, "grad_norm": 0.08470553904771805, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8770780.0, "reward": 0.5808917284011841, "reward_std": 0.3061688244342804, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16178125143051147, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9181250333786011, "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": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 401.5625, "completions/mean_terminated_length": 401.5625, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.1572, "grad_norm": 0.08725372701883316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8781973.0, "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": 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": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 640.125, "completions/mean_terminated_length": 640.125, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.1574, "grad_norm": 0.052062906324863434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8796967.0, "reward": 0.585041880607605, "reward_std": 0.4000430405139923, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1700812429189682, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9662500619888306, "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": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 239.875, "completions/mean_terminated_length": 239.875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.1576, "grad_norm": 0.1986783891916275, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8804389.0, "reward": 1.0667731761932373, "reward_std": 0.5074304938316345, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5710436701774597, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9918124675750732, "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": 874.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 625.0, "completions/mean_terminated_length": 625.0, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1578, "grad_norm": 0.045465972274541855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8818229.0, "reward": 0.528425931930542, "reward_std": 0.23431476950645447, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18185000121593475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8862500190734863, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 839.0, "completions/mean_length": 565.25, "completions/mean_terminated_length": 459.3846435546875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.158, "grad_norm": 0.031974270939826965, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8831577.0, "reward": 0.44023215770721436, "reward_std": 0.21959947049617767, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06796249747276306, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7774999737739563, "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": 538.0, "completions/max_terminated_length": 538.0, "completions/mean_length": 348.1875, "completions/mean_terminated_length": 348.1875, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1582, "grad_norm": 0.07687834650278091, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8840908.0, "reward": 0.6342853307723999, "reward_std": 0.3644215166568756, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.206068754196167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9487500190734863, "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": 1011.0, "completions/mean_length": 619.5, "completions/mean_terminated_length": 592.5333862304688, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.1584, "grad_norm": 0.08318689465522766, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8854692.0, "reward": 0.7207695841789246, "reward_std": 0.5305309295654297, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3790374994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.8268749713897705, "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": 0.0, "completions/max_length": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 409.125, "completions/mean_terminated_length": 409.125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.1586, "grad_norm": 0.04583007097244263, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8864822.0, "reward": 0.5345321297645569, "reward_std": 0.24676057696342468, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1940625011920929, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8531249761581421, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 507.5, "completions/mean_terminated_length": 507.5, "completions/min_length": 383.0, "completions/min_terminated_length": 383.0, "epoch": 0.1588, "grad_norm": 0.07488162070512772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8876878.0, "reward": 1.2080323696136475, "reward_std": 0.44637495279312134, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7285624742507935, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 446.0, "completions/max_terminated_length": 446.0, "completions/mean_length": 284.1875, "completions/mean_terminated_length": 284.1875, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.159, "grad_norm": 0.1393558531999588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8885569.0, "reward": 0.8118727207183838, "reward_std": 0.3427903354167938, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6237437725067139, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.43687501549720764, "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": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 616.625, "completions/mean_terminated_length": 616.625, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.1592, "grad_norm": 0.012046661227941513, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8900235.0, "reward": 0.5914071798324585, "reward_std": 0.07400838285684586, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.18281248211860657, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 410.3125, "completions/mean_terminated_length": 369.4000244140625, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.1594, "grad_norm": 0.10464581102132797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8910816.0, "reward": 0.9842540621757507, "reward_std": 0.5476416945457458, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5310062766075134, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9043750166893005, "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": 398.0, "completions/max_terminated_length": 398.0, "completions/mean_length": 273.0625, "completions/mean_terminated_length": 273.0625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.1596, "grad_norm": 0.15388378500938416, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8919201.0, "reward": 1.134822964668274, "reward_std": 0.4859182834625244, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6446437835693359, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9656250476837158, "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": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 397.1875, "completions/mean_terminated_length": 397.1875, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.1598, "grad_norm": 0.09889975935220718, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8929540.0, "reward": 0.6658353805541992, "reward_std": 0.3256657123565674, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20666874945163727, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568750262260437, "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": 748.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 447.9375, "completions/mean_terminated_length": 447.9375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.16, "grad_norm": 0.01604982651770115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8940435.0, "reward": 0.5829666256904602, "reward_std": 0.09198056161403656, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.16593123972415924, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9056249856948853, "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": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 418.0, "completions/mean_terminated_length": 418.0, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1602, "grad_norm": 0.12285543233156204, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8951507.0, "reward": 0.9895197153091431, "reward_std": 0.5425058603286743, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5415375232696533, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "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": 0.0, "completions/max_length": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 356.875, "completions/mean_terminated_length": 356.875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1604, "grad_norm": 0.08992427587509155, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8961329.0, "reward": 1.2592010498046875, "reward_std": 0.42950257658958435, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7684000134468079, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.966249942779541, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 385.5625, "completions/mean_terminated_length": 385.5625, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.1606, "grad_norm": 0.08758917450904846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8971514.0, "reward": 1.224413514137268, "reward_std": 0.5062663555145264, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7613250017166138, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 759.0, "completions/max_terminated_length": 759.0, "completions/mean_length": 540.0, "completions/mean_terminated_length": 540.0, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.1608, "grad_norm": 0.08468066155910492, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 8983850.0, "reward": 0.7878912687301636, "reward_std": 0.3055386245250702, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5757812261581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.625, "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": 0.0, "completions/max_length": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 277.625, "completions/mean_terminated_length": 277.625, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.161, "grad_norm": 0.1403322070837021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 8992308.0, "reward": 1.0831135511398315, "reward_std": 0.4873530864715576, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.603725016117096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 785.6875, "completions/mean_terminated_length": 751.6428833007812, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.1612, "grad_norm": 0.05520561710000038, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9008959.0, "reward": 0.35850438475608826, "reward_std": 0.44275158643722534, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15450625121593475, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.7987499833106995, "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": 422.0, "completions/max_terminated_length": 422.0, "completions/mean_length": 337.3125, "completions/mean_terminated_length": 337.3125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.1614, "grad_norm": 0.10502692312002182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9018164.0, "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": 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": 0.0, "completions/max_length": 383.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 304.125, "completions/mean_terminated_length": 304.125, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1616, "grad_norm": 0.004495013039559126, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9026534.0, "reward": 0.5474071502685547, "reward_std": 0.017231186851859093, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09481249749660492, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9512500166893005, "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": 518.0, "completions/max_terminated_length": 518.0, "completions/mean_length": 422.6875, "completions/mean_terminated_length": 422.6875, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.1618, "grad_norm": 0.06989524513483047, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9037969.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 651.4375, "completions/mean_terminated_length": 626.6000366210938, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.162, "grad_norm": 0.026355186477303505, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9052168.0, "reward": 0.46859174966812134, "reward_std": 0.18404363095760345, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.062181249260902405, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 0.0, "completions/max_length": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 460.5, "completions/mean_terminated_length": 460.5, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 0.1622, "grad_norm": 0.15066687762737274, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9063792.0, "reward": 0.7404322028160095, "reward_std": 0.5053399205207825, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.29336249828338623, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 433.5625, "completions/mean_terminated_length": 433.5625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.1624, "grad_norm": 0.08488256484270096, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9074505.0, "reward": 0.784507155418396, "reward_std": 0.4261093735694885, "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": 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": 296.0, "completions/max_terminated_length": 296.0, "completions/mean_length": 199.125, "completions/mean_terminated_length": 199.125, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.1626, "grad_norm": 0.10635248571634293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9081163.0, "reward": 0.6863071918487549, "reward_std": 0.3373173773288727, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.24761250615119934, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "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": 0.0, "completions/max_length": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 352.1875, "completions/mean_terminated_length": 352.1875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 0.1628, "grad_norm": 0.06065044924616814, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9090526.0, "reward": 0.5947510004043579, "reward_std": 0.3031080365180969, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18950000405311584, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8981250524520874, "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": 0.0, "completions/max_length": 514.0, "completions/max_terminated_length": 514.0, "completions/mean_length": 418.3125, "completions/mean_terminated_length": 418.3125, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.163, "grad_norm": 0.05003814026713371, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9101203.0, "reward": 0.6137509346008301, "reward_std": 0.2370653748512268, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.16499999165534973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 236.125, "completions/mean_terminated_length": 236.125, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.1632, "grad_norm": 0.09573971480131149, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9108757.0, "reward": 1.4363884925842285, "reward_std": 0.24970568716526031, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9352749586105347, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 481.0, "completions/max_terminated_length": 481.0, "completions/mean_length": 362.25, "completions/mean_terminated_length": 362.25, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.1634, "grad_norm": 0.014846810139715672, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9118345.0, "reward": 0.5654478073120117, "reward_std": 0.0757063627243042, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.130893737077713, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "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": 951.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 608.9375, "completions/mean_terminated_length": 608.9375, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.1636, "grad_norm": 0.13806717097759247, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9132408.0, "reward": 0.6350668668746948, "reward_std": 0.5600976943969727, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2701312303543091, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.9081250429153442, "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": 617.0, "completions/max_terminated_length": 617.0, "completions/mean_length": 414.375, "completions/mean_terminated_length": 414.375, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.1638, "grad_norm": 0.06816283613443375, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9143422.0, "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": 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": 0.0, "completions/max_length": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 431.9375, "completions/mean_terminated_length": 431.9375, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.164, "grad_norm": 0.0041159470565617085, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9154333.0, "reward": 0.5366197824478149, "reward_std": 0.018931839615106583, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07323749363422394, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 541.875, "completions/mean_terminated_length": 541.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1642, "grad_norm": 0.03850904479622841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9166891.0, "reward": 0.7567977905273438, "reward_std": 0.30287250876426697, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.38859373331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8468749523162842, "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": 810.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 535.5, "completions/mean_terminated_length": 535.5, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 0.1644, "grad_norm": 0.0976162850856781, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9179395.0, "reward": 0.953825831413269, "reward_std": 0.5171494483947754, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5326499938964844, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8618749976158142, "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": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 477.5, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.1646, "grad_norm": 0.08378565311431885, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9190843.0, "reward": 1.1655480861663818, "reward_std": 0.4091593623161316, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7685937285423279, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8406250476837158, "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": 374.0, "completions/max_terminated_length": 374.0, "completions/mean_length": 285.0625, "completions/mean_terminated_length": 285.0625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.1648, "grad_norm": 0.08663854748010635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9199484.0, "reward": 1.3784511089324951, "reward_std": 0.33153069019317627, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8819000124931335, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9812500476837158, "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": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 393.0625, "completions/mean_terminated_length": 393.0625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.165, "grad_norm": 0.04111549258232117, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9209917.0, "reward": 0.5641415119171143, "reward_std": 0.19341957569122314, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.19078125059604645, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8156249523162842, "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": 764.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 512.1875, "completions/mean_terminated_length": 512.1875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.1652, "grad_norm": 0.07790876179933548, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9221968.0, "reward": 1.1409791707992554, "reward_std": 0.47769156098365784, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6569562554359436, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 975.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 594.375, "completions/mean_terminated_length": 594.375, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.1654, "grad_norm": 0.07898940145969391, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9235414.0, "reward": 0.989363431930542, "reward_std": 0.5429185628890991, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5412249565124512, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8987499475479126, "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": 605.0, "completions/max_terminated_length": 605.0, "completions/mean_length": 495.3125, "completions/mean_terminated_length": 495.3125, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.1656, "grad_norm": 0.037848811596632004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9247163.0, "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": 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": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 609.9375, "completions/mean_terminated_length": 609.9375, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.1658, "grad_norm": 0.08184965699911118, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9260986.0, "reward": 0.9154696464538574, "reward_std": 0.42504024505615234, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5184375047683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.84375, "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": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 435.25, "completions/mean_terminated_length": 435.25, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.166, "grad_norm": 0.10811331123113632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9271262.0, "reward": 0.748816728591919, "reward_std": 0.4675890803337097, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3101312518119812, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 946.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 374.8125, "completions/mean_terminated_length": 374.8125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.1662, "grad_norm": 0.06483389437198639, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9281707.0, "reward": 0.6734231114387512, "reward_std": 0.33447811007499695, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22184374928474426, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "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": 0.0, "completions/max_length": 291.0, "completions/max_terminated_length": 291.0, "completions/mean_length": 218.25, "completions/mean_terminated_length": 218.25, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 0.1664, "grad_norm": 0.00012332519690971822, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9288655.0, "reward": 1.499860405921936, "reward_std": 0.00030073351808823645, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9997187852859497, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9868749976158142, "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": 947.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 478.875, "completions/mean_terminated_length": 478.875, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.1666, "grad_norm": 0.0745505765080452, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9300765.0, "reward": 0.5717073678970337, "reward_std": 0.2834388315677643, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1434125006198883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 461.0, "completions/max_terminated_length": 461.0, "completions/mean_length": 420.5625, "completions/mean_terminated_length": 420.5625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.1668, "grad_norm": 0.04921198636293411, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9312726.0, "reward": 1.437657356262207, "reward_std": 0.24738894402980804, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9378125071525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9750000238418579, "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": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 511.75, "completions/mean_terminated_length": 511.75, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.167, "grad_norm": 0.06748367100954056, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9325906.0, "reward": 1.2781257629394531, "reward_std": 0.3359594941139221, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8062499761581421, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7749999761581421, "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": 655.0, "completions/max_terminated_length": 655.0, "completions/mean_length": 487.0, "completions/mean_terminated_length": 487.0, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.1672, "grad_norm": 0.09330663830041885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9338162.0, "reward": 0.7975510358810425, "reward_std": 0.4224477708339691, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3450999855995178, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9375, "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": 318.0, "completions/max_terminated_length": 318.0, "completions/mean_length": 213.0, "completions/mean_terminated_length": 213.0, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 0.1674, "grad_norm": 0.09700243920087814, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9345026.0, "reward": 0.6492823362350464, "reward_std": 0.3322223126888275, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17356251180171967, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "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": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 366.3125, "completions/mean_terminated_length": 366.3125, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.1676, "grad_norm": 0.08195086568593979, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9355191.0, "reward": 0.9543445110321045, "reward_std": 0.3782859444618225, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6586874723434448, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6399999856948853, "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": 0.0, "completions/max_length": 660.0, "completions/max_terminated_length": 660.0, "completions/mean_length": 483.375, "completions/mean_terminated_length": 483.375, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.1678, "grad_norm": 0.0829891711473465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9367085.0, "reward": 1.0808321237564087, "reward_std": 0.48977750539779663, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5991624593734741, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 494.9375, "completions/mean_terminated_length": 459.66668701171875, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.168, "grad_norm": 0.0702085867524147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9378492.0, "reward": 0.7059385776519775, "reward_std": 0.4063042402267456, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.28687500953674316, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8125, "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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 348.5, "completions/mean_terminated_length": 348.5, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.1682, "grad_norm": 0.10679551213979721, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9388772.0, "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": 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": 736.0, "completions/max_terminated_length": 736.0, "completions/mean_length": 461.5625, "completions/mean_terminated_length": 461.5625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.1684, "grad_norm": 0.05930584296584129, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9400013.0, "reward": 0.7382041215896606, "reward_std": 0.3307977020740509, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3514062464237213, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.840624988079071, "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": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 416.375, "completions/mean_terminated_length": 416.375, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.1686, "grad_norm": 0.05762471631169319, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9411571.0, "reward": 0.5756257772445679, "reward_std": 0.26385989785194397, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2762500047683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7906249761581421, "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": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 312.0625, "completions/mean_terminated_length": 312.0625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1688, "grad_norm": 0.09009838104248047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9420500.0, "reward": 1.2925729751586914, "reward_std": 0.38213175535202026, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8351437449455261, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8956249952316284, "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": 600.0, "completions/max_terminated_length": 600.0, "completions/mean_length": 437.125, "completions/mean_terminated_length": 437.125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.169, "grad_norm": 0.07817455381155014, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9432742.0, "reward": 1.0111602544784546, "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": 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": 0.0, "completions/max_length": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 406.375, "completions/mean_terminated_length": 406.375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.1692, "grad_norm": 0.06197820231318474, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9443452.0, "reward": 0.6167978048324585, "reward_std": 0.23591399192810059, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17109374701976776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 617.6875, "completions/mean_terminated_length": 559.6428833007812, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.1694, "grad_norm": 0.059814270585775375, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9457415.0, "reward": 0.7517975568771362, "reward_std": 0.4570951759815216, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4410937428474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6968749761581421, "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": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 313.9375, "completions/mean_terminated_length": 313.9375, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.1696, "grad_norm": 0.037678930908441544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9466982.0, "reward": 0.7891412377357483, "reward_std": 0.16720573604106903, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5782812833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5843750238418579, "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": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 471.875, "completions/mean_terminated_length": 471.875, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.1698, "grad_norm": 0.09433849155902863, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9478004.0, "reward": 1.176485300064087, "reward_std": 0.5064594149589539, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.727968692779541, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8531249761581421, "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": 0.0, "completions/max_length": 549.0, "completions/max_terminated_length": 549.0, "completions/mean_length": 398.875, "completions/mean_terminated_length": 398.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.17, "grad_norm": 0.035558927804231644, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9488050.0, "reward": 0.505935549736023, "reward_std": 0.13615022599697113, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07436874508857727, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 342.0, "completions/max_terminated_length": 342.0, "completions/mean_length": 285.25, "completions/mean_terminated_length": 285.25, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.1702, "grad_norm": 0.10638915002346039, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9496358.0, "reward": 0.7118948698043823, "reward_std": 0.3911559581756592, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23628747463226318, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9712499976158142, "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": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 516.5, "completions/mean_terminated_length": 516.5, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.1704, "grad_norm": 0.09492616355419159, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9508990.0, "reward": 0.8746010661125183, "reward_std": 0.4752671420574188, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4991999864578247, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8731250166893005, "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": 505.0, "completions/max_terminated_length": 505.0, "completions/mean_length": 414.25, "completions/mean_terminated_length": 414.25, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.1706, "grad_norm": 0.09312217682600021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9519362.0, "reward": 1.0235165357589722, "reward_std": 0.49033665657043457, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5470312833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 404.125, "completions/mean_terminated_length": 404.125, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.1708, "grad_norm": 0.06735387444496155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9529348.0, "reward": 0.6868758201599121, "reward_std": 0.3232496380805969, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3737500309944153, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.706250011920929, "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": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 587.6875, "completions/mean_terminated_length": 587.6875, "completions/min_length": 433.0, "completions/min_terminated_length": 433.0, "epoch": 0.171, "grad_norm": 0.054488468915224075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9543375.0, "reward": 0.9042944312095642, "reward_std": 0.39196786284446716, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5585874915122986, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7212499380111694, "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": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 462.25, "completions/mean_terminated_length": 462.25, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.1712, "grad_norm": 0.09072792530059814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9554883.0, "reward": 0.9214633107185364, "reward_std": 0.4256834089756012, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5304250121116638, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8062499761581421, "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": 0.0, "completions/max_length": 479.0, "completions/max_terminated_length": 479.0, "completions/mean_length": 345.0625, "completions/mean_terminated_length": 345.0625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.1714, "grad_norm": 0.042226072400808334, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9565940.0, "reward": 0.6009634733200073, "reward_std": 0.2400716096162796, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13942497968673706, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 473.6875, "completions/mean_terminated_length": 437.0000305175781, "completions/min_length": 323.0, "completions/min_terminated_length": 323.0, "epoch": 0.1716, "grad_norm": 0.11844104528427124, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9577743.0, "reward": 1.0117950439453125, "reward_std": 0.5952736735343933, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5860874652862549, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8524999618530273, "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": 700.0, "completions/max_terminated_length": 700.0, "completions/mean_length": 520.6875, "completions/mean_terminated_length": 520.6875, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.1718, "grad_norm": 0.06845036149024963, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9589882.0, "reward": 1.0508599281311035, "reward_std": 0.3920350670814514, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.8517187833786011, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.5562499761581421, "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": 401.0, "completions/max_terminated_length": 401.0, "completions/mean_length": 310.75, "completions/mean_terminated_length": 310.75, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.172, "grad_norm": 0.06992074847221375, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9598198.0, "reward": 0.7256072163581848, "reward_std": 0.3841713070869446, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26371249556541443, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 339.25, "completions/mean_terminated_length": 339.25, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.1722, "grad_norm": 0.13183853030204773, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9608058.0, "reward": 0.8426573276519775, "reward_std": 0.4571767747402191, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37281250953674316, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 0.0, "completions/max_length": 484.0, "completions/max_terminated_length": 484.0, "completions/mean_length": 309.9375, "completions/mean_terminated_length": 309.9375, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1724, "grad_norm": 0.09966348856687546, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9617033.0, "reward": 1.2591261863708496, "reward_std": 0.42945781350135803, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7682499885559082, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 644.625, "completions/mean_terminated_length": 590.4285888671875, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 0.1726, "grad_norm": 0.018820656463503838, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9632115.0, "reward": 0.4559009373188019, "reward_std": 0.17880627512931824, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.03680000081658363, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.856249988079071, "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": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 389.1875, "completions/mean_terminated_length": 389.1875, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.1728, "grad_norm": 0.06054278090596199, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9642198.0, "reward": 0.595994770526886, "reward_std": 0.2414565533399582, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12948749959468842, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "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": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 576.375, "completions/mean_terminated_length": 576.375, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.173, "grad_norm": 0.06664030998945236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9655452.0, "reward": 0.8143478035926819, "reward_std": 0.462262362241745, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.44119375944137573, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.84375, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 520.0625, "completions/mean_terminated_length": 448.0714416503906, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1732, "grad_norm": 0.11004719138145447, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9667997.0, "reward": 0.6666324138641357, "reward_std": 0.504180908203125, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.33326250314712524, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7387500405311584, "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": 0.0, "completions/max_length": 954.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 575.625, "completions/mean_terminated_length": 575.625, "completions/min_length": 373.0, "completions/min_terminated_length": 373.0, "epoch": 0.1734, "grad_norm": 0.08163341134786606, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9680951.0, "reward": 0.8678914308547974, "reward_std": 0.5111504197120667, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4232812523841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.859375, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 642.625, "completions/mean_terminated_length": 617.2000122070312, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 0.1736, "grad_norm": 0.03297848999500275, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9695905.0, "reward": 0.6230478286743164, "reward_std": 0.24840590357780457, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.30859375, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7093750238418579, "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": 0.0, "completions/max_length": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 357.3125, "completions/mean_terminated_length": 357.3125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.1738, "grad_norm": 0.08611465245485306, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9705590.0, "reward": 1.201875925064087, "reward_std": 0.45477786660194397, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7162500619888306, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "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": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 284.875, "completions/mean_terminated_length": 284.875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.174, "grad_norm": 0.12535490095615387, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9714340.0, "reward": 1.0800542831420898, "reward_std": 0.491043359041214, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5976062417030334, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "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": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 392.1875, "completions/mean_terminated_length": 392.1875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.1742, "grad_norm": 0.1196524053812027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9724359.0, "reward": 0.9597508907318115, "reward_std": 0.5170716047286987, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5444999933242798, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8762500286102295, "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": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 355.8125, "completions/mean_terminated_length": 355.8125, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.1744, "grad_norm": 0.034785427153110504, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9734932.0, "reward": 1.4395947456359863, "reward_std": 0.23755879700183868, "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": 0.9575000405311584, "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": 0.0, "completions/max_length": 698.0, "completions/max_terminated_length": 698.0, "completions/mean_length": 372.5625, "completions/mean_terminated_length": 372.5625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1746, "grad_norm": 0.08641352504491806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9745661.0, "reward": 0.7406008243560791, "reward_std": 0.3871920108795166, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4187000095844269, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6974999904632568, "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": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 352.5625, "completions/mean_terminated_length": 352.5625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.1748, "grad_norm": 0.10685385018587112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9755734.0, "reward": 1.0442726612091064, "reward_std": 0.47621846199035645, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5885437726974487, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9256250262260437, "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": 879.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 539.875, "completions/mean_terminated_length": 539.875, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 0.175, "grad_norm": 0.03879769146442413, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9769860.0, "reward": 0.4841262102127075, "reward_std": 0.2087613344192505, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09324999898672104, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9412500262260437, "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": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 330.6875, "completions/mean_terminated_length": 330.6875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 0.1752, "grad_norm": 0.1221831738948822, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9779247.0, "reward": 0.8972604870796204, "reward_std": 0.48178839683532715, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.41951873898506165, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 0.0, "completions/max_length": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 410.9375, "completions/mean_terminated_length": 410.9375, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1754, "grad_norm": 0.07581642270088196, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9789950.0, "reward": 0.7238291501998901, "reward_std": 0.3846590220928192, "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": 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": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 387.875, "completions/mean_terminated_length": 387.875, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.1756, "grad_norm": 0.0729639008641243, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9800428.0, "reward": 1.3184853792190552, "reward_std": 0.3883984088897705, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.824468731880188, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618750214576721, "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": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 443.6875, "completions/mean_terminated_length": 443.6875, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.1758, "grad_norm": 0.06142503768205643, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9813063.0, "reward": 0.7878129482269287, "reward_std": 0.2903765141963959, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6381250619888306, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.4124999940395355, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 472.0, "completions/mean_length": 358.875, "completions/mean_terminated_length": 314.5333557128906, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.176, "grad_norm": 0.15027356147766113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9822725.0, "reward": 0.8707039952278137, "reward_std": 0.520106852054596, "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": 0.890625, "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": 373.0, "completions/max_terminated_length": 373.0, "completions/mean_length": 231.1875, "completions/mean_terminated_length": 231.1875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.1762, "grad_norm": 0.10998479276895523, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9829896.0, "reward": 1.3762603998184204, "reward_std": 0.33625534176826477, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8775187730789185, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9706249833106995, "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": 390.0, "completions/max_terminated_length": 390.0, "completions/mean_length": 301.3125, "completions/mean_terminated_length": 301.3125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1764, "grad_norm": 0.08399257063865662, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9838653.0, "reward": 1.3780760765075684, "reward_std": 0.331384539604187, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8811500072479248, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "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": 720.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 476.1875, "completions/mean_terminated_length": 476.1875, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.1766, "grad_norm": 0.12600408494472504, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9849856.0, "reward": 0.9413292407989502, "reward_std": 0.5612711310386658, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5701562166213989, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8156249523162842, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 793.6875, "completions/mean_terminated_length": 740.5385131835938, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 0.1768, "grad_norm": 0.04133396968245506, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9867755.0, "reward": 0.679375410079956, "reward_std": 0.37514007091522217, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5462499856948853, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.3687500059604645, "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": 0.0, "completions/max_length": 218.0, "completions/max_terminated_length": 218.0, "completions/mean_length": 153.375, "completions/mean_terminated_length": 153.375, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 0.177, "grad_norm": 0.14505985379219055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9873857.0, "reward": 0.712941586971283, "reward_std": 0.39025869965553284, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2383812516927719, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 424.0, "completions/max_terminated_length": 424.0, "completions/mean_length": 369.1875, "completions/mean_terminated_length": 369.1875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.1772, "grad_norm": 0.13976198434829712, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9884228.0, "reward": 0.9877102971076965, "reward_std": 0.5441197752952576, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.537918746471405, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 1022.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 604.1875, "completions/mean_terminated_length": 604.1875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.1774, "grad_norm": 0.06904175877571106, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9897975.0, "reward": 1.03733229637146, "reward_std": 0.44486844539642334, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.637162446975708, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8425000309944153, "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": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 297.5, "completions/mean_terminated_length": 297.5, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1776, "grad_norm": 0.11505410075187683, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9907199.0, "reward": 0.8377290964126587, "reward_std": 0.46056145429611206, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3629562556743622, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 0.0, "completions/max_length": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 415.3125, "completions/mean_terminated_length": 415.3125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.1778, "grad_norm": 0.09187395125627518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9917620.0, "reward": 0.983254075050354, "reward_std": 0.473599910736084, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5290062427520752, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9343750476837158, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 776.9375, "completions/mean_terminated_length": 664.6364135742188, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.178, "grad_norm": 0.049554675817489624, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9935427.0, "reward": 0.38507866859436035, "reward_std": 0.368903249502182, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.20765624940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.5343749523162842, "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": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 398.625, "completions/mean_terminated_length": 398.625, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.1782, "grad_norm": 0.07707927376031876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9946061.0, "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": 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": 0.0, "completions/max_length": 679.0, "completions/max_terminated_length": 679.0, "completions/mean_length": 344.125, "completions/mean_terminated_length": 344.125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1784, "grad_norm": 0.048206888139247894, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9955871.0, "reward": 1.4042823314666748, "reward_std": 0.20393167436122894, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.9960625171661377, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7950000166893005, "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": 0.0, "completions/max_length": 681.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 471.6875, "completions/mean_terminated_length": 471.6875, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.1786, "grad_norm": 0.093772754073143, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9967194.0, "reward": 1.0963762998580933, "reward_std": 0.5506523251533508, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6302499771118164, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 949.0, "completions/mean_length": 790.5625, "completions/mean_terminated_length": 736.6923217773438, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.1788, "grad_norm": 0.04563399404287338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 9983507.0, "reward": 0.6835944652557373, "reward_std": 0.46086186170578003, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4296875, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5687500238418579, "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": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 440.375, "completions/mean_terminated_length": 440.375, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.179, "grad_norm": 0.08562420308589935, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 9994697.0, "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": 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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 705.625, "completions/mean_terminated_length": 632.1538696289062, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.1792, "grad_norm": 0.06486500054597855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10009907.0, "reward": 0.5948445200920105, "reward_std": 0.4668413996696472, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.31468749046325684, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.65625, "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": 330.0, "completions/max_terminated_length": 330.0, "completions/mean_length": 271.3125, "completions/mean_terminated_length": 271.3125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.1794, "grad_norm": 0.06931557506322861, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10018840.0, "reward": 0.5815012454986572, "reward_std": 0.2456277459859848, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10050000250339508, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 856.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 604.125, "completions/mean_terminated_length": 604.125, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.1796, "grad_norm": 0.07968532294034958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10032106.0, "reward": 1.0178073644638062, "reward_std": 0.49726545810699463, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5356125235557556, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 322.0, "completions/max_terminated_length": 322.0, "completions/mean_length": 279.875, "completions/mean_terminated_length": 279.875, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.1798, "grad_norm": 0.1934324949979782, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10040392.0, "reward": 1.3457696437835693, "reward_std": 0.43006205558776855, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8790374994277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 300.4375, "completions/mean_terminated_length": 300.4375, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.18, "grad_norm": 0.139454185962677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10048927.0, "reward": 1.0076324939727783, "reward_std": 0.5080369114875793, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5152624845504761, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9774999618530273, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 683.0625, "completions/mean_terminated_length": 660.3333740234375, "completions/min_length": 390.0, "completions/min_terminated_length": 390.0, "epoch": 0.1802, "grad_norm": 0.07276476174592972, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10064528.0, "reward": 0.961250901222229, "reward_std": 0.5230851769447327, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5475000143051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8312499523162842, "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": 707.0, "completions/max_terminated_length": 707.0, "completions/mean_length": 540.1875, "completions/mean_terminated_length": 540.1875, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.1804, "grad_norm": 0.08355804532766342, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10077395.0, "reward": 1.0706260204315186, "reward_std": 0.5018928050994873, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5787500143051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 320.9375, "completions/mean_terminated_length": 320.9375, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.1806, "grad_norm": 0.1325538605451584, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10086194.0, "reward": 0.7316413521766663, "reward_std": 0.38882797956466675, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6507812738418579, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.31562501192092896, "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": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 645.1875, "completions/mean_terminated_length": 645.1875, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.1808, "grad_norm": 0.059810489416122437, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10100101.0, "reward": 0.8344539999961853, "reward_std": 0.372050017118454, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.48140624165534973, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7656249403953552, "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": 735.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 461.0, "completions/mean_terminated_length": 461.0, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.181, "grad_norm": 0.08399044722318649, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10111445.0, "reward": 0.6957821846008301, "reward_std": 0.3641204237937927, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.3290625214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8718750476837158, "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": 0.0, "completions/max_length": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 347.6875, "completions/mean_terminated_length": 347.6875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.1812, "grad_norm": 0.10336150228977203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10120976.0, "reward": 1.0804543495178223, "reward_std": 0.49077853560447693, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5984062552452087, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 0.0, "completions/max_length": 421.0, "completions/max_terminated_length": 421.0, "completions/mean_length": 278.625, "completions/mean_terminated_length": 278.625, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 0.1814, "grad_norm": 0.10438327491283417, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10129418.0, "reward": 0.7839040756225586, "reward_std": 0.42654669284820557, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3178062438964844, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9556249976158142, "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": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 292.625, "completions/mean_terminated_length": 292.625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.1816, "grad_norm": 0.11367963999509811, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10138100.0, "reward": 0.7814011573791504, "reward_std": 0.42866191267967224, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31279999017715454, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "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": 389.0, "completions/max_terminated_length": 389.0, "completions/mean_length": 345.3125, "completions/mean_terminated_length": 345.3125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.1818, "grad_norm": 0.109011210501194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10147977.0, "reward": 1.286144733428955, "reward_std": 0.47103166580200195, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8222874999046326, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.953125, "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": 0.0, "completions/max_length": 581.0, "completions/max_terminated_length": 581.0, "completions/mean_length": 331.625, "completions/mean_terminated_length": 331.625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.182, "grad_norm": 0.00041244772728532553, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10157155.0, "reward": 1.4992952346801758, "reward_std": 0.0012870528735220432, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9985874891281128, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9775000214576721, "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": 367.0, "completions/max_terminated_length": 367.0, "completions/mean_length": 305.875, "completions/mean_terminated_length": 305.875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.1822, "grad_norm": 0.10767611861228943, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10165841.0, "reward": 0.9672665596008301, "reward_std": 0.4842281639575958, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4970312714576721, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 627.0, "completions/max_terminated_length": 627.0, "completions/mean_length": 351.3125, "completions/mean_terminated_length": 351.3125, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.1824, "grad_norm": 0.1470964550971985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10175142.0, "reward": 0.6565510034561157, "reward_std": 0.4565841555595398, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25060001015663147, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 539.0, "completions/max_terminated_length": 539.0, "completions/mean_length": 437.25, "completions/mean_terminated_length": 437.25, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 0.1826, "grad_norm": 0.08727588504552841, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10186234.0, "reward": 0.7836697101593018, "reward_std": 0.4268513321876526, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3173374831676483, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 0.0, "completions/max_length": 764.0, "completions/max_terminated_length": 764.0, "completions/mean_length": 446.5, "completions/mean_terminated_length": 446.5, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.1828, "grad_norm": 0.0348476842045784, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10196946.0, "reward": 0.5169887542724609, "reward_std": 0.17076410353183746, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09647499769926071, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8774999976158142, "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": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 377.0625, "completions/mean_terminated_length": 377.0625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.183, "grad_norm": 0.03391481190919876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10207363.0, "reward": 0.626794695854187, "reward_std": 0.18560890853405, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1910874992609024, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8462499976158142, "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": 567.0, "completions/max_terminated_length": 567.0, "completions/mean_length": 379.375, "completions/mean_terminated_length": 379.375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 0.1832, "grad_norm": 0.03315068036317825, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10217369.0, "reward": 0.5819541811943054, "reward_std": 0.11835138499736786, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1639062464237213, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8843749761581421, "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": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 262.25, "completions/mean_terminated_length": 262.25, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 0.1834, "grad_norm": 0.12843763828277588, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10225677.0, "reward": 1.0101540088653564, "reward_std": 0.41210898756980896, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6453062295913696, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6681250333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 548.1875, "completions/mean_terminated_length": 516.4666748046875, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.1836, "grad_norm": 0.08686204999685287, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10238336.0, "reward": 1.0413137674331665, "reward_std": 0.5513089895248413, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5826250314712524, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9137500524520874, "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": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 367.1875, "completions/mean_terminated_length": 367.1875, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.1838, "grad_norm": 0.10994183272123337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10248099.0, "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": 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": 0.0, "completions/max_length": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 496.4375, "completions/mean_terminated_length": 496.4375, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.184, "grad_norm": 0.09770074486732483, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10260106.0, "reward": 1.1355698108673096, "reward_std": 0.48515743017196655, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6461374759674072, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9687500596046448, "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": 375.0, "completions/max_terminated_length": 375.0, "completions/mean_length": 256.25, "completions/mean_terminated_length": 256.25, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.1842, "grad_norm": 0.11500829458236694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10268094.0, "reward": 1.097732424736023, "reward_std": 0.47756442427635193, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6329625248908997, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 616.0, "completions/mean_terminated_length": 588.800048828125, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 0.1844, "grad_norm": 0.056221429258584976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10281918.0, "reward": 0.8002229928970337, "reward_std": 0.4335375130176544, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4754437804222107, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6706249713897705, "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": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 461.8125, "completions/mean_terminated_length": 461.8125, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1846, "grad_norm": 0.1284475475549698, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10293499.0, "reward": 1.105385422706604, "reward_std": 0.5402451753616333, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6482687592506409, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 466.0, "completions/max_terminated_length": 466.0, "completions/mean_length": 409.375, "completions/mean_terminated_length": 409.375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.1848, "grad_norm": 0.0563986711204052, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10304177.0, "reward": 0.6584166288375854, "reward_std": 0.3284234404563904, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19183124601840973, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 742.3125, "completions/mean_terminated_length": 573.2999877929688, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.185, "grad_norm": 0.03805384412407875, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10321830.0, "reward": 0.33992254734039307, "reward_std": 0.2722000479698181, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.05484374985098839, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.5968749523162842, "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": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 260.0, "completions/mean_terminated_length": 260.0, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.1852, "grad_norm": 0.09844627976417542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10331334.0, "reward": 1.3158917427062988, "reward_std": 0.3951994478702545, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8192812204360962, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "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": 416.0, "completions/max_terminated_length": 416.0, "completions/mean_length": 280.375, "completions/mean_terminated_length": 280.375, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 0.1854, "grad_norm": 0.005910719279199839, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10339372.0, "reward": 0.5336228609085083, "reward_std": 0.018026428297162056, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06724375486373901, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "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": 0.0, "completions/max_length": 628.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 447.4375, "completions/mean_terminated_length": 447.4375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1856, "grad_norm": 0.11822158098220825, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10350307.0, "reward": 0.8350011110305786, "reward_std": 0.49195945262908936, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41999998688697815, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.918749988079071, "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": 943.0, "completions/mean_length": 666.6875, "completions/mean_terminated_length": 642.86669921875, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.1858, "grad_norm": 0.08759641647338867, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10364990.0, "reward": 0.8091883659362793, "reward_std": 0.528496503829956, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4308750033378601, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8462499976158142, "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": 0.0, "completions/max_length": 457.0, "completions/max_terminated_length": 457.0, "completions/mean_length": 354.875, "completions/mean_terminated_length": 354.875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.186, "grad_norm": 0.045584242790937424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10374332.0, "reward": 1.439657211303711, "reward_std": 0.23757553100585938, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9418125152587891, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 532.6875, "completions/mean_terminated_length": 532.6875, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 0.1862, "grad_norm": 0.05608801171183586, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10387207.0, "reward": 0.634610652923584, "reward_std": 0.3679763972759247, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20671874284744263, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.949999988079071, "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": 0.0, "completions/max_length": 361.0, "completions/max_terminated_length": 361.0, "completions/mean_length": 257.5, "completions/mean_terminated_length": 257.5, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.1864, "grad_norm": 0.05468692630529404, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10395583.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 545.0, "completions/mean_length": 409.125, "completions/mean_terminated_length": 368.13336181640625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.1866, "grad_norm": 0.09394679963588715, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10405793.0, "reward": 0.689454197883606, "reward_std": 0.42435353994369507, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25390625, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 654.0, "completions/mean_terminated_length": 629.3333740234375, "completions/min_length": 413.0, "completions/min_terminated_length": 413.0, "epoch": 0.1868, "grad_norm": 0.03829975053668022, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10420785.0, "reward": 0.6953129768371582, "reward_std": 0.3078993558883667, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.515625, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.518750011920929, "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": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 431.4375, "completions/mean_terminated_length": 431.4375, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.187, "grad_norm": 0.017333636060357094, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10431512.0, "reward": 0.586250901222229, "reward_std": 0.09303884953260422, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17249998450279236, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.0, "completions/max_length": 355.0, "completions/max_terminated_length": 355.0, "completions/mean_length": 260.3125, "completions/mean_terminated_length": 260.3125, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.1872, "grad_norm": 0.16305580735206604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10439725.0, "reward": 1.1384072303771973, "reward_std": 0.4808678925037384, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6518125534057617, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 396.0, "completions/max_terminated_length": 396.0, "completions/mean_length": 264.125, "completions/mean_terminated_length": 264.125, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.1874, "grad_norm": 0.1034621000289917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10447919.0, "reward": 0.6803138256072998, "reward_std": 0.3413779139518738, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2356249988079071, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 721.3125, "completions/mean_terminated_length": 620.4166870117188, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.1876, "grad_norm": 0.10848738998174667, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10463828.0, "reward": 0.6521099805831909, "reward_std": 0.5854042768478394, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.36671876907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6437499523162842, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 519.0, "completions/mean_length": 431.625, "completions/mean_terminated_length": 392.13336181640625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.1878, "grad_norm": 0.10551110655069351, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10475294.0, "reward": 0.8392977714538574, "reward_std": 0.49426478147506714, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4285937547683716, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.840624988079071, "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": 376.0, "completions/max_terminated_length": 376.0, "completions/mean_length": 285.125, "completions/mean_terminated_length": 285.125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.188, "grad_norm": 0.10155215859413147, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10483552.0, "reward": 0.6638821363449097, "reward_std": 0.32604631781578064, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20276249945163727, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 763.3125, "completions/mean_terminated_length": 726.0714721679688, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.1882, "grad_norm": 0.03317964822053909, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10499957.0, "reward": 0.5850006341934204, "reward_std": 0.26845091581344604, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.29499998688697815, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6875, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 859.5625, "completions/mean_terminated_length": 784.8181762695312, "completions/min_length": 445.0, "completions/min_terminated_length": 445.0, "epoch": 0.1884, "grad_norm": 0.0448501817882061, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10517966.0, "reward": 0.43351632356643677, "reward_std": 0.33713918924331665, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1795312464237213, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5343749523162842, "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": 876.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 575.5625, "completions/mean_terminated_length": 575.5625, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.1886, "grad_norm": 0.07828065752983093, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10531063.0, "reward": 1.251929521560669, "reward_std": 0.442489892244339, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7538562417030334, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756249785423279, "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": 860.0, "completions/max_terminated_length": 860.0, "completions/mean_length": 529.4375, "completions/mean_terminated_length": 529.4375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.1888, "grad_norm": 0.04971790313720703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10544094.0, "reward": 0.6684383749961853, "reward_std": 0.31416791677474976, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.33687499165534973, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8156249523162842, "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": 0.0, "completions/max_length": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 491.125, "completions/mean_terminated_length": 491.125, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.189, "grad_norm": 0.07145471125841141, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10556000.0, "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": 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": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 432.625, "completions/mean_terminated_length": 432.625, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.1892, "grad_norm": 0.0732908695936203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10566506.0, "reward": 1.3207042217254639, "reward_std": 0.38299617171287537, "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": 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": 0.0, "completions/max_length": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 234.6875, "completions/mean_terminated_length": 234.6875, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.1894, "grad_norm": 0.13023148477077484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10573829.0, "reward": 0.7045825719833374, "reward_std": 0.3951305150985718, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2216624915599823, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 480.0, "completions/max_terminated_length": 480.0, "completions/mean_length": 331.1875, "completions/mean_terminated_length": 331.1875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.1896, "grad_norm": 0.10131102055311203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10582984.0, "reward": 1.3215514421463013, "reward_std": 0.3834165036678314, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8305999636650085, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 433.0, "completions/max_terminated_length": 433.0, "completions/mean_length": 311.5, "completions/mean_terminated_length": 311.5, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.1898, "grad_norm": 0.09628794342279434, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10591984.0, "reward": 1.3210792541503906, "reward_std": 0.3831825852394104, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8296562433242798, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9668749570846558, "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": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 484.0, "completions/mean_terminated_length": 484.0, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.19, "grad_norm": 0.10416998714208603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10603712.0, "reward": 1.0524227619171143, "reward_std": 0.4602947533130646, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6048437356948853, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "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": 0.0, "completions/max_length": 630.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 554.9375, "completions/mean_terminated_length": 554.9375, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.1902, "grad_norm": 0.09461748600006104, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10616223.0, "reward": 1.22703218460083, "reward_std": 0.5014948844909668, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7665625214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.953125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 521.3125, "completions/mean_terminated_length": 487.8000183105469, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.1904, "grad_norm": 0.09911958128213882, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10628676.0, "reward": 1.1625009775161743, "reward_std": 0.5255565047264099, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.699999988079071, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8812500238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 606.9375, "completions/mean_terminated_length": 579.1333618164062, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.1906, "grad_norm": 0.06509513407945633, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10643635.0, "reward": 1.10257887840271, "reward_std": 0.49194884300231934, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7051562666893005, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7281249761581421, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 642.1875, "completions/mean_terminated_length": 587.6428833007812, "completions/min_length": 365.0, "completions/min_terminated_length": 365.0, "epoch": 0.1908, "grad_norm": 0.026392797008156776, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10657926.0, "reward": 0.4765383303165436, "reward_std": 0.1862787902355194, "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": 0.8349999785423279, "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": 0.0, "completions/max_length": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 457.3125, "completions/mean_terminated_length": 457.3125, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.191, "grad_norm": 0.018512237817049026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10669195.0, "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": 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": 0.0, "completions/max_length": 485.0, "completions/max_terminated_length": 485.0, "completions/mean_length": 337.5625, "completions/mean_terminated_length": 337.5625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.1912, "grad_norm": 0.0031696599908173084, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10678500.0, "reward": 0.5445165634155273, "reward_std": 0.011709402315318584, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 649.8125, "completions/mean_terminated_length": 624.86669921875, "completions/min_length": 366.0, "completions/min_terminated_length": 366.0, "epoch": 0.1914, "grad_norm": 0.05327887460589409, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10693105.0, "reward": 0.6830697059631348, "reward_std": 0.4265739917755127, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.24113750457763672, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.90625, "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": 633.0, "completions/max_terminated_length": 633.0, "completions/mean_length": 474.5, "completions/mean_terminated_length": 474.5, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 0.1916, "grad_norm": 0.10706768184900284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10705369.0, "reward": 0.9582822322845459, "reward_std": 0.5252807140350342, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.541562557220459, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8937499523162842, "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": 0.0, "completions/max_length": 692.0, "completions/max_terminated_length": 692.0, "completions/mean_length": 440.125, "completions/mean_terminated_length": 440.125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1918, "grad_norm": 0.04786478728055954, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10717099.0, "reward": 0.9122661352157593, "reward_std": 0.24248665571212769, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6995312571525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5093749761581421, "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": 710.0, "completions/max_terminated_length": 710.0, "completions/mean_length": 518.25, "completions/mean_terminated_length": 518.25, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.192, "grad_norm": 0.06900861859321594, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10729919.0, "reward": 0.8971761465072632, "reward_std": 0.4817030429840088, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4193500280380249, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612499475479126, "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": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 449.5, "completions/mean_terminated_length": 449.5, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.1922, "grad_norm": 0.09004713594913483, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10741095.0, "reward": 0.8407759666442871, "reward_std": 0.45833417773246765, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.369049996137619, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 947.0, "completions/mean_length": 541.5625, "completions/mean_terminated_length": 509.4000244140625, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 0.1924, "grad_norm": 0.11204483360052109, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10753840.0, "reward": 0.9279696941375732, "reward_std": 0.5065914392471313, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5434374809265137, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "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": 517.0, "completions/max_terminated_length": 517.0, "completions/mean_length": 433.625, "completions/mean_terminated_length": 433.625, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.1926, "grad_norm": 0.10456161946058273, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10764570.0, "reward": 1.3457822799682617, "reward_std": 0.4300668239593506, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8790625333786011, "rewards/confidence_one_or_zero": 0.0625, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 726.4375, "completions/mean_terminated_length": 657.7692260742188, "completions/min_length": 446.0, "completions/min_terminated_length": 446.0, "epoch": 0.1928, "grad_norm": 0.057726018130779266, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10779969.0, "reward": 0.621016263961792, "reward_std": 0.3506935238838196, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4295312166213989, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.49062496423721313, "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": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 345.9375, "completions/mean_terminated_length": 345.9375, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.193, "grad_norm": 0.15156497061252594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10789488.0, "reward": 1.140075922012329, "reward_std": 0.47832077741622925, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6551500558853149, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 384.0, "completions/max_terminated_length": 384.0, "completions/mean_length": 290.3125, "completions/mean_terminated_length": 290.3125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 0.1932, "grad_norm": 0.08652031421661377, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10797893.0, "reward": 0.5532541871070862, "reward_std": 0.28468459844589233, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.10650623589754105, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9743499755859375, "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": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 389.375, "completions/mean_terminated_length": 389.375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.1934, "grad_norm": 0.11526899039745331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10808411.0, "reward": 0.9927853345870972, "reward_std": 0.47418251633644104, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5480687618255615, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 485.5625, "completions/mean_terminated_length": 449.66668701171875, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.1936, "grad_norm": 0.051155850291252136, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10819732.0, "reward": 0.5612508654594421, "reward_std": 0.32150888442993164, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1850000023841858, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8624999523162842, "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": 0.0, "completions/max_length": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 264.3125, "completions/mean_terminated_length": 264.3125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.1938, "grad_norm": 0.1611718237400055, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10827545.0, "reward": 1.0114669799804688, "reward_std": 0.504432201385498, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5229312181472778, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806250333786011, "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": 285.0, "completions/max_terminated_length": 285.0, "completions/mean_length": 245.0, "completions/mean_terminated_length": 245.0, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 0.194, "grad_norm": 0.16325311362743378, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10835577.0, "reward": 0.8359947800636292, "reward_std": 0.4619424641132355, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.35948750376701355, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "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": 0.0, "completions/max_length": 328.0, "completions/max_terminated_length": 328.0, "completions/mean_length": 249.8125, "completions/mean_terminated_length": 249.8125, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.1942, "grad_norm": 0.2800917625427246, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10843366.0, "reward": 1.178279161453247, "reward_std": 0.5111371874809265, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7315562963485718, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9418750405311584, "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": 0.0, "completions/max_length": 397.0, "completions/max_terminated_length": 397.0, "completions/mean_length": 273.1875, "completions/mean_terminated_length": 273.1875, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.1944, "grad_norm": 0.07670574635267258, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10851497.0, "reward": 0.6283603310585022, "reward_std": 0.24658355116844177, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.19421875476837158, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "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": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 305.6875, "completions/mean_terminated_length": 305.6875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.1946, "grad_norm": 0.07581884413957596, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10860356.0, "reward": 0.6199136972427368, "reward_std": 0.36848771572113037, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.17732499539852142, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9668750166893005, "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": 367.0, "completions/max_terminated_length": 367.0, "completions/mean_length": 323.0625, "completions/mean_terminated_length": 323.0625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.1948, "grad_norm": 0.10387760400772095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10869205.0, "reward": 0.8397228717803955, "reward_std": 0.4593709409236908, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.36694374680519104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 680.0, "completions/max_terminated_length": 680.0, "completions/mean_length": 394.8125, "completions/mean_terminated_length": 394.8125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.195, "grad_norm": 0.009155089035630226, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10880018.0, "reward": 0.5543758869171143, "reward_std": 0.022499974817037582, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.10875000059604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 559.0, "completions/max_terminated_length": 559.0, "completions/mean_length": 396.625, "completions/mean_terminated_length": 396.625, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.1952, "grad_norm": 0.10947235673666, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10890012.0, "reward": 1.0608478784561157, "reward_std": 0.46332868933677673, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6216937303543091, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8956249952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 449.125, "completions/mean_terminated_length": 410.8000183105469, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.1954, "grad_norm": 0.09842311590909958, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10901486.0, "reward": 0.6920197606086731, "reward_std": 0.41816872358322144, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.25903749465942383, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8737500309944153, "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": 901.0, "completions/max_terminated_length": 901.0, "completions/mean_length": 572.375, "completions/mean_terminated_length": 572.375, "completions/min_length": 393.0, "completions/min_terminated_length": 393.0, "epoch": 0.1956, "grad_norm": 0.04906604439020157, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10914852.0, "reward": 0.6575663089752197, "reward_std": 0.24716679751873016, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3776312470436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6893750429153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 596.3125, "completions/mean_terminated_length": 567.800048828125, "completions/min_length": 338.0, "completions/min_terminated_length": 338.0, "epoch": 0.1958, "grad_norm": 0.07807519286870956, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10928441.0, "reward": 0.6716293096542358, "reward_std": 0.46526601910591125, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.28075623512268066, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8675000071525574, "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": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 352.6875, "completions/mean_terminated_length": 352.6875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 0.196, "grad_norm": 0.08645819872617722, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10938004.0, "reward": 0.665847897529602, "reward_std": 0.325700044631958, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20669375360012054, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 743.75, "completions/mean_terminated_length": 703.7142944335938, "completions/min_length": 482.0, "completions/min_terminated_length": 482.0, "epoch": 0.1962, "grad_norm": 0.05107646808028221, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10954272.0, "reward": 0.7192975282669067, "reward_std": 0.38007238507270813, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5010937452316284, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5218750238418579, "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": 602.0, "completions/max_terminated_length": 602.0, "completions/mean_length": 399.0625, "completions/mean_terminated_length": 399.0625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.1964, "grad_norm": 0.05540488287806511, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10965057.0, "reward": 0.6144540309906006, "reward_std": 0.3033972978591919, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22890624403953552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8218749761581421, "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": 418.0, "completions/max_terminated_length": 418.0, "completions/mean_length": 348.25, "completions/mean_terminated_length": 348.25, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.1966, "grad_norm": 0.08775684237480164, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 10974837.0, "reward": 1.4055665731430054, "reward_std": 0.37481793761253357, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9361312389373779, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 477.5, "completions/mean_terminated_length": 477.5, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 0.1968, "grad_norm": 0.0999101847410202, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 10986525.0, "reward": 1.0644947290420532, "reward_std": 0.46602663397789, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6289874911308289, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8899999856948853, "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": 0.0, "completions/max_length": 940.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 630.6875, "completions/mean_terminated_length": 630.6875, "completions/min_length": 408.0, "completions/min_terminated_length": 408.0, "epoch": 0.197, "grad_norm": 0.07081470638513565, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11000920.0, "reward": 0.9846729040145874, "reward_std": 0.47760048508644104, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5318437814712524, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9231249690055847, "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": 595.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 467.875, "completions/mean_terminated_length": 467.875, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.1972, "grad_norm": 0.09319306910037994, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11012390.0, "reward": 0.9948571920394897, "reward_std": 0.4729340970516205, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5522124767303467, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 546.8125, "completions/mean_terminated_length": 546.8125, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.1974, "grad_norm": 0.08070841431617737, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11026083.0, "reward": 0.7519540786743164, "reward_std": 0.46550777554512024, "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": 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": 520.0, "completions/max_terminated_length": 520.0, "completions/mean_length": 406.0625, "completions/mean_terminated_length": 406.0625, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 0.1976, "grad_norm": 0.0981731191277504, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11036116.0, "reward": 1.073201060295105, "reward_std": 0.49922576546669006, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5839000344276428, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9725000262260437, "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": 513.0, "completions/max_terminated_length": 513.0, "completions/mean_length": 392.1875, "completions/mean_terminated_length": 392.1875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.1978, "grad_norm": 0.07702351361513138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11046615.0, "reward": 1.3160417079925537, "reward_std": 0.39363473653793335, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8195812702178955, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631250500679016, "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": 370.0, "completions/max_terminated_length": 370.0, "completions/mean_length": 308.4375, "completions/mean_terminated_length": 308.4375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.198, "grad_norm": 0.12271591275930405, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11055502.0, "reward": 1.3184230327606201, "reward_std": 0.3883673846721649, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.824343740940094, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 0.0, "completions/max_length": 493.0, "completions/max_terminated_length": 493.0, "completions/mean_length": 340.125, "completions/mean_terminated_length": 340.125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.1982, "grad_norm": 0.05734725296497345, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11064848.0, "reward": 0.5559264421463013, "reward_std": 0.21820910274982452, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.17435069382190704, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7945112586021423, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 585.9375, "completions/mean_terminated_length": 556.7333374023438, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.1984, "grad_norm": 0.05583760514855385, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11078207.0, "reward": 0.5694540739059448, "reward_std": 0.3061487674713135, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20140624046325684, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6843750476837158, "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": 0.0, "completions/max_length": 377.0, "completions/max_terminated_length": 377.0, "completions/mean_length": 240.625, "completions/mean_terminated_length": 240.625, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 0.1986, "grad_norm": 0.11313647031784058, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11087017.0, "reward": 1.3744919300079346, "reward_std": 0.34136638045310974, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.873981237411499, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750071525574, "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": 601.0, "completions/max_terminated_length": 601.0, "completions/mean_length": 423.6875, "completions/mean_terminated_length": 423.6875, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.1988, "grad_norm": 0.10671141743659973, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11097476.0, "reward": 1.1398417949676514, "reward_std": 0.4795105457305908, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6546812653541565, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 769.0, "completions/mean_length": 673.875, "completions/mean_terminated_length": 650.5333862304688, "completions/min_length": 492.0, "completions/min_terminated_length": 492.0, "epoch": 0.199, "grad_norm": 0.045101143419742584, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11112706.0, "reward": 0.7732038497924805, "reward_std": 0.36173170804977417, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.48390626907348633, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6968749761581421, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 498.75, "completions/mean_terminated_length": 423.71429443359375, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.1992, "grad_norm": 0.06976167857646942, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11124782.0, "reward": 0.7863286733627319, "reward_std": 0.3459542393684387, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6351562738418579, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.27812498807907104, "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": 437.0, "completions/max_terminated_length": 437.0, "completions/mean_length": 353.8125, "completions/mean_terminated_length": 353.8125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.1994, "grad_norm": 0.11561089754104614, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11134203.0, "reward": 0.7790133953094482, "reward_std": 0.42934203147888184, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3080250024795532, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 639.5625, "completions/mean_terminated_length": 639.5625, "completions/min_length": 531.0, "completions/min_terminated_length": 531.0, "epoch": 0.1996, "grad_norm": 0.06360184401273727, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11148436.0, "reward": 0.910313606262207, "reward_std": 0.4724206030368805, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.445624977350235, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.949999988079071, "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": 0.0, "completions/max_length": 583.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 352.75, "completions/mean_terminated_length": 352.75, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.1998, "grad_norm": 0.10347910225391388, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11158016.0, "reward": 0.474729061126709, "reward_std": 0.1856698840856552, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07445624470710754, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9574999809265137, "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": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 214.4375, "completions/mean_terminated_length": 214.4375, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.2, "grad_norm": 0.1931922286748886, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11165287.0, "reward": 1.3459852933883667, "reward_std": 0.4301441013813019, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8794687986373901, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 580.125, "completions/mean_terminated_length": 550.5333862304688, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 0.2002, "grad_norm": 0.10192502290010452, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11178121.0, "reward": 0.7423854470252991, "reward_std": 0.4706040024757385, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.29726874828338623, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9056249856948853, "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": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 346.3125, "completions/mean_terminated_length": 346.3125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.2004, "grad_norm": 0.05828890576958656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11187854.0, "reward": 0.5989855527877808, "reward_std": 0.240728497505188, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1354687511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 283.0, "completions/max_terminated_length": 283.0, "completions/mean_length": 218.625, "completions/mean_terminated_length": 218.625, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 0.2006, "grad_norm": 0.17129772901535034, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11195208.0, "reward": 1.0178886651992798, "reward_std": 0.49734944105148315, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5357750058174133, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9649999737739563, "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": 744.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 561.4375, "completions/mean_terminated_length": 561.4375, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.2008, "grad_norm": 0.0407516285777092, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11209423.0, "reward": 0.5748043656349182, "reward_std": 0.24715831875801086, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08710625022649765, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9868749976158142, "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": 0.0, "completions/max_length": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 564.9375, "completions/mean_terminated_length": 564.9375, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.201, "grad_norm": 0.060166534036397934, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11222318.0, "reward": 0.9994537234306335, "reward_std": 0.3409086763858795, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.7489062547683716, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.515625, "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": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 513.375, "completions/mean_terminated_length": 513.375, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.2012, "grad_norm": 0.0028128328267484903, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11235316.0, "reward": 0.5432791709899902, "reward_std": 0.015061733312904835, "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": 0.9556249976158142, "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": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 878.8125, "completions/mean_terminated_length": 791.7000122070312, "completions/min_length": 472.0, "completions/min_terminated_length": 472.0, "epoch": 0.2014, "grad_norm": 0.05326048284769058, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11253553.0, "reward": 0.3601568043231964, "reward_std": 0.4332830011844635, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15781250596046448, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.5625, "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": 393.0, "completions/max_terminated_length": 393.0, "completions/mean_length": 268.4375, "completions/mean_terminated_length": 268.4375, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.2016, "grad_norm": 0.10795660316944122, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11262184.0, "reward": 1.1041417121887207, "reward_std": 0.42258772253990173, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6457812786102295, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.809374988079071, "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": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 598.0625, "completions/mean_terminated_length": 598.0625, "completions/min_length": 412.0, "completions/min_terminated_length": 412.0, "epoch": 0.2018, "grad_norm": 0.030057625845074654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11276505.0, "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": 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": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 654.5, "completions/mean_terminated_length": 654.5, "completions/min_length": 443.0, "completions/min_terminated_length": 443.0, "epoch": 0.202, "grad_norm": 0.05001161992549896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11291201.0, "reward": 0.6709915399551392, "reward_std": 0.28093379735946655, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.40448126196861267, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6256250143051147, "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": 558.0, "completions/max_terminated_length": 558.0, "completions/mean_length": 342.0, "completions/mean_terminated_length": 342.0, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.2022, "grad_norm": 0.1346791386604309, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11301329.0, "reward": 1.0644478797912598, "reward_std": 0.46553218364715576, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6288937330245972, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8893749713897705, "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": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 723.125, "completions/mean_terminated_length": 489.1111145019531, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.2024, "grad_norm": 0.10852913558483124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11318003.0, "reward": 0.7607036828994751, "reward_std": 0.6823482513427734, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.45890623331069946, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.484375, "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": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 356.4375, "completions/mean_terminated_length": 356.4375, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.2026, "grad_norm": 0.11467701196670532, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11327578.0, "reward": 0.9519981145858765, "reward_std": 0.49905121326446533, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.46649378538131714, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750071525574, "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": 405.0, "completions/max_terminated_length": 405.0, "completions/mean_length": 222.75, "completions/mean_terminated_length": 222.75, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.2028, "grad_norm": 0.05400991812348366, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11335798.0, "reward": 0.945625901222229, "reward_std": 0.1335521936416626, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.8912500143051147, "rewards/confidence_one_or_zero": 0.75, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.15000000596046448, "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": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 338.0, "completions/mean_terminated_length": 338.0, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.203, "grad_norm": 0.006658186204731464, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11345814.0, "reward": 0.5366135835647583, "reward_std": 0.019282974302768707, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.0732249915599823, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 0.0, "completions/max_length": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 488.3125, "completions/mean_terminated_length": 488.3125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 0.2032, "grad_norm": 0.12053482234477997, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11357291.0, "reward": 0.7902981042861938, "reward_std": 0.49478238821029663, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3305937647819519, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9418749809265137, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 584.375, "completions/mean_terminated_length": 555.0667114257812, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 0.2034, "grad_norm": 0.046091724187135696, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11371009.0, "reward": 0.7107819318771362, "reward_std": 0.2650849223136902, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4840624928474426, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.543749988079071, "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": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 310.3125, "completions/mean_terminated_length": 310.3125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.2036, "grad_norm": 0.18722420930862427, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11379670.0, "reward": 0.7845821380615234, "reward_std": 0.4262435734272003, "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": 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": 784.0, "completions/mean_length": 544.3125, "completions/mean_terminated_length": 512.3333740234375, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.2038, "grad_norm": 0.05511815845966339, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11391867.0, "reward": 0.8453909158706665, "reward_std": 0.268689900636673, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7532812356948853, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.31562501192092896, "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": 691.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 425.875, "completions/mean_terminated_length": 425.875, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.204, "grad_norm": 0.07996446639299393, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11402233.0, "reward": 0.7110135555267334, "reward_std": 0.33662089705467224, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.42202502489089966, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6962499618530273, "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": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 439.125, "completions/mean_terminated_length": 439.125, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.2042, "grad_norm": 0.07519666105508804, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11413227.0, "reward": 1.3657822608947754, "reward_std": 0.33156800270080566, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8565624952316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9187500476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 613.25, "completions/mean_terminated_length": 585.86669921875, "completions/min_length": 437.0, "completions/min_terminated_length": 437.0, "epoch": 0.2044, "grad_norm": 0.05362611636519432, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11426975.0, "reward": 0.6028915047645569, "reward_std": 0.2909277677536011, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2057812511920929, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8531249761581421, "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": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 357.0625, "completions/mean_terminated_length": 357.0625, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.2046, "grad_norm": 0.11441509425640106, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11436304.0, "reward": 0.7748104333877563, "reward_std": 0.43286439776420593, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.2996187210083008, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9693750143051147, "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": 420.0, "completions/max_terminated_length": 420.0, "completions/mean_length": 313.4375, "completions/mean_terminated_length": 313.4375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.2048, "grad_norm": 0.08498985320329666, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11445575.0, "reward": 0.71176677942276, "reward_std": 0.39114412665367126, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.23603123426437378, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 0.0, "completions/max_length": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 266.25, "completions/mean_terminated_length": 266.25, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 0.205, "grad_norm": 0.10608658194541931, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11453531.0, "reward": 0.6572978496551514, "reward_std": 0.3290681540966034, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.18959373235702515, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 670.0, "completions/max_terminated_length": 670.0, "completions/mean_length": 409.5625, "completions/mean_terminated_length": 409.5625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.2052, "grad_norm": 0.07935056835412979, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11463876.0, "reward": 0.6618759632110596, "reward_std": 0.2789709270000458, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.38625001907348633, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6625000238418579, "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": 0.0, "completions/max_length": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 317.75, "completions/mean_terminated_length": 317.75, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.2054, "grad_norm": 0.0867161750793457, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11472992.0, "reward": 0.5442855358123779, "reward_std": 0.18648526072502136, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15106874704360962, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9081249833106995, "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": 690.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 550.6875, "completions/mean_terminated_length": 550.6875, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.2056, "grad_norm": 0.057341307401657104, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11485611.0, "reward": 1.339141607284546, "reward_std": 0.3452208936214447, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8657812476158142, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281250238418579, "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": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 861.4375, "completions/mean_terminated_length": 735.0, "completions/min_length": 490.0, "completions/min_terminated_length": 490.0, "epoch": 0.2058, "grad_norm": 0.05317949131131172, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11503378.0, "reward": 0.4709910750389099, "reward_std": 0.4918883442878723, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.3169812560081482, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.33812499046325684, "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": 297.0, "completions/max_terminated_length": 297.0, "completions/mean_length": 233.8125, "completions/mean_terminated_length": 233.8125, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 0.206, "grad_norm": 0.00018998404266312718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11511455.0, "reward": 1.4993321895599365, "reward_std": 0.0006022164016030729, "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": 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": 536.0, "completions/max_terminated_length": 536.0, "completions/mean_length": 397.125, "completions/mean_terminated_length": 397.125, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.2062, "grad_norm": 0.09650295972824097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11521857.0, "reward": 0.7428134083747864, "reward_std": 0.37628814578056335, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2981249988079071, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.925000011920929, "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": 0.0, "completions/max_length": 453.0, "completions/max_terminated_length": 453.0, "completions/mean_length": 296.375, "completions/mean_terminated_length": 296.375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.2064, "grad_norm": 0.0020974765066057444, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11530727.0, "reward": 0.99648517370224, "reward_std": 0.008958701975643635, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.992968738079071, "rewards/confidence_one_or_zero": 0.8125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.03437500074505806, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 707.875, "completions/mean_terminated_length": 634.923095703125, "completions/min_length": 375.0, "completions/min_terminated_length": 375.0, "epoch": 0.2066, "grad_norm": 0.048152729868888855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11546293.0, "reward": 0.6623442769050598, "reward_std": 0.39741864800453186, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.44968751072883606, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.550000011920929, "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": 495.0, "completions/max_terminated_length": 495.0, "completions/mean_length": 369.4375, "completions/mean_terminated_length": 369.4375, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.2068, "grad_norm": 0.08784278482198715, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11557436.0, "reward": 0.7804259657859802, "reward_std": 0.42903488874435425, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31084999442100525, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9637500047683716, "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": 561.0, "completions/max_terminated_length": 561.0, "completions/mean_length": 424.5625, "completions/mean_terminated_length": 424.5625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.207, "grad_norm": 0.0912984162569046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11568421.0, "reward": 0.7518635392189026, "reward_std": 0.46601593494415283, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31622499227523804, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 0.0, "completions/max_length": 585.0, "completions/max_terminated_length": 585.0, "completions/mean_length": 404.8125, "completions/mean_terminated_length": 404.8125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.2072, "grad_norm": 0.007143941707909107, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11578882.0, "reward": 0.5495197772979736, "reward_std": 0.027453087270259857, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09903749823570251, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9487500190734863, "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": 0.0, "completions/max_length": 487.0, "completions/max_terminated_length": 487.0, "completions/mean_length": 334.5, "completions/mean_terminated_length": 334.5, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 0.2074, "grad_norm": 0.10754523426294327, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11588186.0, "reward": 1.1983509063720703, "reward_std": 0.45705369114875793, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7092000246047974, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9325000047683716, "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": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 447.625, "completions/mean_terminated_length": 447.625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.2076, "grad_norm": 0.042405690997838974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11599380.0, "reward": 0.6002352833747864, "reward_std": 0.13411292433738708, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2004687488079071, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.859375, "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": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 564.0, "completions/mean_terminated_length": 564.0, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.2078, "grad_norm": 0.08387491106987, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11612372.0, "reward": 1.0831916332244873, "reward_std": 0.4867900311946869, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6038812398910522, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9518749713897705, "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": 0.0, "completions/max_length": 752.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 523.1875, "completions/mean_terminated_length": 523.1875, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 0.208, "grad_norm": 0.09543872624635696, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11625559.0, "reward": 0.8589011430740356, "reward_std": 0.4568343758583069, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.40529999136924744, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 679.25, "completions/mean_terminated_length": 656.2667236328125, "completions/min_length": 405.0, "completions/min_terminated_length": 405.0, "epoch": 0.2082, "grad_norm": 0.05978613719344139, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11640507.0, "reward": 1.0543947219848633, "reward_std": 0.5400896668434143, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6087875366210938, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8962500095367432, "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": 529.0, "completions/max_terminated_length": 529.0, "completions/mean_length": 401.3125, "completions/mean_terminated_length": 401.3125, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.2084, "grad_norm": 0.09345707297325134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11651056.0, "reward": 1.1394541263580322, "reward_std": 0.47920018434524536, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6539062261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "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": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 443.6875, "completions/mean_terminated_length": 443.6875, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 0.2086, "grad_norm": 0.09928780794143677, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11661867.0, "reward": 0.8977073431015015, "reward_std": 0.4816831648349762, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.42041251063346863, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 471.75, "completions/mean_terminated_length": 434.933349609375, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.2088, "grad_norm": 0.09947839379310608, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11673495.0, "reward": 1.2177354097366333, "reward_std": 0.48168161511421204, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.8104687333106995, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7906250357627869, "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": 373.0, "completions/max_terminated_length": 373.0, "completions/mean_length": 312.0625, "completions/mean_terminated_length": 312.0625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.209, "grad_norm": 0.00010883739741984755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11682824.0, "reward": 1.4989041090011597, "reward_std": 0.00041871648863889277, "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": 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": 931.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 492.6875, "completions/mean_terminated_length": 492.6875, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 0.2092, "grad_norm": 0.03714122250676155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11694483.0, "reward": 0.7035163640975952, "reward_std": 0.15253134071826935, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.40703123807907104, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.734375, "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": 0.0, "completions/max_length": 712.0, "completions/max_terminated_length": 712.0, "completions/mean_length": 516.625, "completions/mean_terminated_length": 516.625, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.2094, "grad_norm": 0.06700491160154343, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11706877.0, "reward": 1.0009384155273438, "reward_std": 0.4123528003692627, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6268749833106995, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.75, "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": 371.0, "completions/max_terminated_length": 371.0, "completions/mean_length": 293.9375, "completions/mean_terminated_length": 293.9375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.2096, "grad_norm": 0.042349737137556076, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11715612.0, "reward": 1.4627102613449097, "reward_std": 0.1447230875492096, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.9879187345504761, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9193750023841858, "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": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 635.5625, "completions/mean_terminated_length": 635.5625, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.2098, "grad_norm": 0.03071289137005806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11731141.0, "reward": 0.8335943222045898, "reward_std": 0.20218414068222046, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.667187511920929, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.42499998211860657, "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": 204.0, "completions/max_terminated_length": 204.0, "completions/mean_length": 166.8125, "completions/mean_terminated_length": 166.8125, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.21, "grad_norm": 0.00033903439179994166, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11737298.0, "reward": 1.4996670484542847, "reward_std": 0.000548550917301327, "rewards/accuracy_reward": 1.0, "rewards/brier_reward": 0.9993312358856201, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.984375, "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": 204.0, "completions/max_terminated_length": 204.0, "completions/mean_length": 163.5625, "completions/mean_terminated_length": 163.5625, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 0.2102, "grad_norm": 0.23725832998752594, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11743643.0, "reward": 1.197469711303711, "reward_std": 0.462647020816803, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7074375152587891, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 608.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 470.0, "completions/mean_terminated_length": 470.0, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 0.2104, "grad_norm": 0.05658648908138275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11756107.0, "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": 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": 0.0, "completions/max_length": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 439.0, "completions/mean_terminated_length": 439.0, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.2106, "grad_norm": 0.0025226410944014788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11766683.0, "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": 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": 0.0, "completions/max_length": 490.0, "completions/max_terminated_length": 490.0, "completions/mean_length": 367.1875, "completions/mean_terminated_length": 367.1875, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.2108, "grad_norm": 0.00311546609736979, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11776446.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 530.0, "completions/mean_length": 381.3125, "completions/mean_terminated_length": 338.4666748046875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.211, "grad_norm": 0.1300138235092163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11786099.0, "reward": 0.8576451539993286, "reward_std": 0.5302474498748779, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4027875065803528, "rewards/confidence_one_or_zero": 0.5, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9225000143051147, "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": 0.0, "completions/max_length": 709.0, "completions/max_terminated_length": 709.0, "completions/mean_length": 487.4375, "completions/mean_terminated_length": 487.4375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 0.2112, "grad_norm": 0.006059120874851942, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11798330.0, "reward": 0.5423449277877808, "reward_std": 0.029583055526018143, "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": 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": 445.0, "completions/max_terminated_length": 445.0, "completions/mean_length": 310.6875, "completions/mean_terminated_length": 310.6875, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.2114, "grad_norm": 0.14617563784122467, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11807605.0, "reward": 0.8396886587142944, "reward_std": 0.45954272150993347, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3668749928474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 361.0625, "completions/mean_terminated_length": 361.0625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.2116, "grad_norm": 0.13125421106815338, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11817606.0, "reward": 1.1110166311264038, "reward_std": 0.533393383026123, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6595312356948853, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9468749761581421, "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": 563.0, "completions/max_terminated_length": 563.0, "completions/mean_length": 387.5625, "completions/mean_terminated_length": 387.5625, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.2118, "grad_norm": 0.12351041287183762, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11827487.0, "reward": 1.0538760423660278, "reward_std": 0.4649668037891388, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6077500581741333, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9275000095367432, "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": 259.0, "completions/max_terminated_length": 259.0, "completions/mean_length": 187.5, "completions/mean_terminated_length": 187.5, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "epoch": 0.212, "grad_norm": 0.11939407140016556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11834263.0, "reward": 0.6997638940811157, "reward_std": 0.2931289076805115, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.337024986743927, "rewards/confidence_one_or_zero": 0.5625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.783750057220459, "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": 0.3125, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 769.375, "completions/mean_terminated_length": 653.6364135742188, "completions/min_length": 508.0, "completions/min_terminated_length": 508.0, "epoch": 0.2122, "grad_norm": 0.0452033169567585, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11850989.0, "reward": 0.4878912568092346, "reward_std": 0.41672083735466003, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22578124701976776, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.5843750238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 787.0, "completions/mean_length": 531.1875, "completions/mean_terminated_length": 498.3333740234375, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 0.2124, "grad_norm": 0.04578615725040436, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11863184.0, "reward": 0.578907310962677, "reward_std": 0.28467220067977905, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.15781250596046448, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.887499988079071, "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": 494.0, "completions/max_terminated_length": 494.0, "completions/mean_length": 351.3125, "completions/mean_terminated_length": 351.3125, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.2126, "grad_norm": 0.07794436812400818, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11872613.0, "reward": 0.6657571792602539, "reward_std": 0.32545217871665955, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2065124809741974, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 632.0, "completions/max_terminated_length": 632.0, "completions/mean_length": 396.75, "completions/mean_terminated_length": 396.75, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.2128, "grad_norm": 0.06401216983795166, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11882977.0, "reward": 1.3477418422698975, "reward_std": 0.3410520851612091, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8829812407493591, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9168750047683716, "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": 943.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 678.0625, "completions/mean_terminated_length": 678.0625, "completions/min_length": 478.0, "completions/min_terminated_length": 478.0, "epoch": 0.213, "grad_norm": 0.0596466138958931, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11897634.0, "reward": 0.999938428401947, "reward_std": 0.46665456891059875, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5623749494552612, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.90625, "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": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 498.125, "completions/mean_terminated_length": 498.125, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.2132, "grad_norm": 0.10157269984483719, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11909444.0, "reward": 0.8408601880073547, "reward_std": 0.45118749141693115, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.49421876668930054, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.765625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 904.0, "completions/mean_length": 665.25, "completions/mean_terminated_length": 641.3333740234375, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.2134, "grad_norm": 0.01426136214286089, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11924552.0, "reward": 0.5342977046966553, "reward_std": 0.15779748558998108, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13109374046325684, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8656249642372131, "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": 727.0, "completions/max_terminated_length": 727.0, "completions/mean_length": 475.875, "completions/mean_terminated_length": 475.875, "completions/min_length": 328.0, "completions/min_terminated_length": 328.0, "epoch": 0.2136, "grad_norm": 0.03438108041882515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11936102.0, "reward": 0.5417665243148804, "reward_std": 0.18712444603443146, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1460312455892563, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9043750166893005, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 374.5625, "completions/mean_terminated_length": 331.2666931152344, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 0.2138, "grad_norm": 0.12470054626464844, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11945887.0, "reward": 0.6025540828704834, "reward_std": 0.36938369274139404, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.392606258392334, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5600000023841858, "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": 435.0, "completions/max_terminated_length": 435.0, "completions/mean_length": 305.0, "completions/mean_terminated_length": 305.0, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.214, "grad_norm": 0.09146949648857117, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11954639.0, "reward": 1.3810011148452759, "reward_std": 0.32379618287086487, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.8869999647140503, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96875, "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": 902.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 553.5, "completions/mean_terminated_length": 553.5, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 0.2142, "grad_norm": 0.015443596057593822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11968279.0, "reward": 0.5496731996536255, "reward_std": 0.08968500047922134, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.09934374690055847, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418749809265137, "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": 777.0, "completions/mean_length": 495.375, "completions/mean_terminated_length": 460.13336181640625, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 0.2144, "grad_norm": 0.0493987500667572, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 11980765.0, "reward": 0.9124224185943604, "reward_std": 0.35025373101234436, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.7623437643051147, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.3968750238418579, "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": 832.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 502.9375, "completions/mean_terminated_length": 502.9375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.2146, "grad_norm": 0.03724708408117294, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 11992524.0, "reward": 0.5586727857589722, "reward_std": 0.15841494500637054, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1798437535762787, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.0, "completions/max_length": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 608.0, "completions/mean_terminated_length": 608.0, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2148, "grad_norm": 0.07942706346511841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12005836.0, "reward": 0.8027228116989136, "reward_std": 0.41738444566726685, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3554437458515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "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": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 424.4375, "completions/mean_terminated_length": 424.4375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 0.215, "grad_norm": 0.028602711856365204, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12016531.0, "reward": 0.5691415071487427, "reward_std": 0.08156237751245499, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.13828125596046448, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "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": 367.0, "completions/max_terminated_length": 367.0, "completions/mean_length": 288.0, "completions/mean_terminated_length": 288.0, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.2152, "grad_norm": 0.04885973036289215, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12026003.0, "reward": 0.5887634754180908, "reward_std": 0.24327810108661652, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.11502499878406525, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 543.5, "completions/mean_terminated_length": 511.4667053222656, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.2154, "grad_norm": 0.06585212051868439, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12039115.0, "reward": 0.5832511186599731, "reward_std": 0.40029677748680115, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1665000021457672, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8525000214576721, "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": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 532.375, "completions/mean_terminated_length": 532.375, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.2156, "grad_norm": 0.08247610926628113, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12051313.0, "reward": 1.323435664176941, "reward_std": 0.37840506434440613, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8343687057495117, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 0.0, "completions/max_length": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 496.1875, "completions/mean_terminated_length": 496.1875, "completions/min_length": 394.0, "completions/min_terminated_length": 394.0, "epoch": 0.2158, "grad_norm": 0.07933402061462402, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12063668.0, "reward": 0.8183477520942688, "reward_std": 0.4159384071826935, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3866937458515167, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9081249833106995, "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": 0.0, "completions/max_length": 483.0, "completions/max_terminated_length": 483.0, "completions/mean_length": 340.375, "completions/mean_terminated_length": 340.375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.216, "grad_norm": 0.09410855174064636, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12073082.0, "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": 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": 502.0, "completions/max_terminated_length": 502.0, "completions/mean_length": 418.125, "completions/mean_terminated_length": 418.125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.2162, "grad_norm": 0.10337626188993454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12083724.0, "reward": 1.1977821588516235, "reward_std": 0.4614655077457428, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7080625295639038, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 359.75, "completions/mean_terminated_length": 359.75, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.2164, "grad_norm": 0.07247496396303177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12093528.0, "reward": 1.3174227476119995, "reward_std": 0.34084352850914, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8848437666893005, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "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": 547.0, "completions/max_terminated_length": 547.0, "completions/mean_length": 358.1875, "completions/mean_terminated_length": 358.1875, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 0.2166, "grad_norm": 0.11051058769226074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12103339.0, "reward": 1.0413291454315186, "reward_std": 0.42788293957710266, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6451562643051147, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.765625, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 596.0625, "completions/mean_terminated_length": 596.0625, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 0.2168, "grad_norm": 0.05957230553030968, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12116908.0, "reward": 1.0181257724761963, "reward_std": 0.35474419593811035, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.7237499952316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5936875343322754, "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": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 522.0, "completions/mean_terminated_length": 522.0, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.217, "grad_norm": 0.1256575882434845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12128892.0, "reward": 0.7892165780067444, "reward_std": 0.4234178960323334, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32843124866485596, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9493750333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 530.25, "completions/mean_terminated_length": 497.3333740234375, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 0.2172, "grad_norm": 0.07057727873325348, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12141488.0, "reward": 0.6689009070396423, "reward_std": 0.34110069274902344, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.337799996137619, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7050000429153442, "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": 320.0, "completions/max_terminated_length": 320.0, "completions/mean_length": 226.75, "completions/mean_terminated_length": 226.75, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.2174, "grad_norm": 0.21957913041114807, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12148700.0, "reward": 1.1326980590820312, "reward_std": 0.4899684190750122, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6403937339782715, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831250309944153, "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": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 386.9375, "completions/mean_terminated_length": 386.9375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.2176, "grad_norm": 0.0716945230960846, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12159067.0, "reward": 0.6541322469711304, "reward_std": 0.20750172436237335, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.30826249718666077, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7200000286102295, "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": 891.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 554.3125, "completions/mean_terminated_length": 554.3125, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.2178, "grad_norm": 0.03614964708685875, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12171872.0, "reward": 0.5912478566169739, "reward_std": 0.20854148268699646, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.24499374628067017, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8431249856948853, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 583.0, "completions/mean_length": 501.8125, "completions/mean_terminated_length": 427.21429443359375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 0.218, "grad_norm": 0.09901876002550125, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12184381.0, "reward": 0.6203914880752563, "reward_std": 0.4026881456375122, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2407812476158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8031250238418579, "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": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 584.5625, "completions/mean_terminated_length": 584.5625, "completions/min_length": 429.0, "completions/min_terminated_length": 429.0, "epoch": 0.2182, "grad_norm": 0.12069027125835419, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12197766.0, "reward": 1.28446364402771, "reward_std": 0.47551947832107544, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8189250230789185, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9731249809265137, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 660.5625, "completions/mean_terminated_length": 636.3333740234375, "completions/min_length": 453.0, "completions/min_terminated_length": 453.0, "epoch": 0.2184, "grad_norm": 0.058833781629800797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12212895.0, "reward": 0.7122414708137512, "reward_std": 0.4235732853412628, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29948124289512634, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8631249666213989, "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": 535.0, "completions/max_terminated_length": 535.0, "completions/mean_length": 336.9375, "completions/mean_terminated_length": 336.9375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.2186, "grad_norm": 0.14645352959632874, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12221822.0, "reward": 0.7188165187835693, "reward_std": 0.4209974408149719, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3126312494277954, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9175000190734863, "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": 889.0, "completions/mean_length": 562.9375, "completions/mean_terminated_length": 532.2000122070312, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.2188, "grad_norm": 0.06188097968697548, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12235101.0, "reward": 0.5646886229515076, "reward_std": 0.2836303412914276, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.12937499582767487, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.375, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 849.5, "completions/mean_terminated_length": 744.7999877929688, "completions/min_length": 502.0, "completions/min_terminated_length": 502.0, "epoch": 0.219, "grad_norm": 0.05175764486193657, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12253125.0, "reward": 0.5315629839897156, "reward_std": 0.514775276184082, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.31312501430511475, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.625, "rewards/mean_confidence_reward": 0.48125001788139343, "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": 612.0, "completions/max_terminated_length": 612.0, "completions/mean_length": 465.9375, "completions/mean_terminated_length": 465.9375, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 0.2192, "grad_norm": 0.06053846701979637, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12264308.0, "reward": 1.32328200340271, "reward_std": 0.2700456380844116, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8965624570846558, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.71875, "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": 542.0, "completions/max_terminated_length": 542.0, "completions/mean_length": 386.75, "completions/mean_terminated_length": 386.75, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.2194, "grad_norm": 0.10728181898593903, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12274400.0, "reward": 0.9595853090286255, "reward_std": 0.49140286445617676, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48166877031326294, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9568749666213989, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 703.75, "completions/mean_terminated_length": 658.0, "completions/min_length": 451.0, "completions/min_terminated_length": 451.0, "epoch": 0.2196, "grad_norm": 0.04643351212143898, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12289708.0, "reward": 0.6390631794929504, "reward_std": 0.40560653805732727, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.40312498807907104, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5375000238418579, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 796.0, "completions/mean_length": 598.9375, "completions/mean_terminated_length": 500.8461608886719, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.2198, "grad_norm": 0.07067551463842392, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12303323.0, "reward": 0.6110941767692566, "reward_std": 0.4044324457645416, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.47218751907348633, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.3812500238418579, "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": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 447.9375, "completions/mean_terminated_length": 447.9375, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 0.22, "grad_norm": 0.09582222998142242, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12314506.0, "reward": 1.009766697883606, "reward_std": 0.48987340927124023, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.51953125, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9281249642372131, "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": 546.0, "completions/max_terminated_length": 546.0, "completions/mean_length": 454.625, "completions/mean_terminated_length": 454.625, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.2202, "grad_norm": 0.07463745027780533, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12326052.0, "reward": 0.9677321910858154, "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": 0.9462500214576721, "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": 400.0, "completions/max_terminated_length": 400.0, "completions/mean_length": 305.5625, "completions/mean_terminated_length": 305.5625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.2204, "grad_norm": 0.15562379360198975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12334989.0, "reward": 1.1397292613983154, "reward_std": 0.47942036390304565, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6544562578201294, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 0.0, "completions/max_length": 454.0, "completions/max_terminated_length": 454.0, "completions/mean_length": 306.375, "completions/mean_terminated_length": 306.375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2206, "grad_norm": 0.06383821368217468, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12343763.0, "reward": 0.6002728343009949, "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": 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": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 424.0625, "completions/mean_terminated_length": 424.0625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 0.2208, "grad_norm": 0.013436906039714813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12354500.0, "reward": 0.5714852809906006, "reward_std": 0.0681101456284523, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.14296874403953552, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "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": 510.0, "completions/max_terminated_length": 510.0, "completions/mean_length": 401.0625, "completions/mean_terminated_length": 401.0625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.221, "grad_norm": 0.11375223845243454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12366501.0, "reward": 0.9825790524482727, "reward_std": 0.4696531593799591, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5276561975479126, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "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": 247.0, "completions/max_terminated_length": 247.0, "completions/mean_length": 200.125, "completions/mean_terminated_length": 200.125, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.2212, "grad_norm": 0.010552164167165756, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12373607.0, "reward": 0.5292887687683105, "reward_std": 0.02303863875567913, "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": 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": 419.0, "completions/max_terminated_length": 419.0, "completions/mean_length": 340.25, "completions/mean_terminated_length": 340.25, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.2214, "grad_norm": 0.12147720903158188, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12383275.0, "reward": 1.0832571983337402, "reward_std": 0.4868498146533966, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6040124893188477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 340.0, "completions/max_terminated_length": 340.0, "completions/mean_length": 241.125, "completions/mean_terminated_length": 241.125, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 0.2216, "grad_norm": 0.18231937289237976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12390781.0, "reward": 0.9527729749679565, "reward_std": 0.49863535165786743, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4680437445640564, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9806249737739563, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 450.5, "completions/mean_terminated_length": 412.2666931152344, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2218, "grad_norm": 0.03566202148795128, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12403141.0, "reward": 0.6025007963180542, "reward_std": 0.1938062608242035, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.26749998331069946, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.78125, "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": 0.0, "completions/max_length": 667.0, "completions/max_terminated_length": 667.0, "completions/mean_length": 361.9375, "completions/mean_terminated_length": 361.9375, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.222, "grad_norm": 0.13760006427764893, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12413172.0, "reward": 1.020782232284546, "reward_std": 0.493945837020874, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.541562557220459, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 957.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 755.75, "completions/mean_terminated_length": 755.75, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.2222, "grad_norm": 0.022711288183927536, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12429760.0, "reward": 0.7784382104873657, "reward_std": 0.14880450069904327, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5568749904632568, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6000000238418579, "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": 325.0, "completions/max_terminated_length": 325.0, "completions/mean_length": 251.75, "completions/mean_terminated_length": 251.75, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 0.2224, "grad_norm": 0.09223799407482147, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12437468.0, "reward": 0.5072165727615356, "reward_std": 0.13586200773715973, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.07693124562501907, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 580.875, "completions/mean_terminated_length": 551.3333740234375, "completions/min_length": 388.0, "completions/min_terminated_length": 388.0, "epoch": 0.2226, "grad_norm": 0.04701581224799156, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12451178.0, "reward": 0.7328789234161377, "reward_std": 0.37865057587623596, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.40325623750686646, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7268750667572021, "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": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 553.8125, "completions/mean_terminated_length": 553.8125, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.2228, "grad_norm": 0.09524925798177719, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12464439.0, "reward": 0.974363386631012, "reward_std": 0.4448559582233429, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5737249851226807, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8174999952316284, "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": 662.0, "completions/max_terminated_length": 662.0, "completions/mean_length": 432.75, "completions/mean_terminated_length": 432.75, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 0.223, "grad_norm": 0.10054317861795425, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12475667.0, "reward": 0.8455479741096497, "reward_std": 0.45533713698387146, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3785937428474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "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": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 434.75, "completions/mean_terminated_length": 434.75, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 0.2232, "grad_norm": 0.11128243803977966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12486895.0, "reward": 0.7750540971755981, "reward_std": 0.4693925678730011, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3626062273979187, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8493750095367432, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 640.0, "completions/mean_length": 458.8125, "completions/mean_terminated_length": 378.0714416503906, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 0.2234, "grad_norm": 0.05092392489314079, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12497724.0, "reward": 0.5394509434700012, "reward_std": 0.3172978162765503, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14139999449253082, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8337500095367432, "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": 0.0, "completions/max_length": 606.0, "completions/max_terminated_length": 606.0, "completions/mean_length": 399.625, "completions/mean_terminated_length": 399.625, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.2236, "grad_norm": 0.1282501369714737, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12508230.0, "reward": 0.8514071702957153, "reward_std": 0.4510110914707184, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 0.0, "completions/max_length": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 533.375, "completions/mean_terminated_length": 533.375, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.2238, "grad_norm": 0.012595482170581818, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12520780.0, "reward": 0.9253910779953003, "reward_std": 0.07102330029010773, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.850781261920929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.3343750238418579, "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": 578.0, "completions/max_terminated_length": 578.0, "completions/mean_length": 496.625, "completions/mean_terminated_length": 496.625, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.224, "grad_norm": 0.0, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12533974.0, "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": 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": 684.0, "completions/max_terminated_length": 684.0, "completions/mean_length": 458.5625, "completions/mean_terminated_length": 458.5625, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 0.2242, "grad_norm": 0.020485449582338333, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12546479.0, "reward": 0.6517977118492126, "reward_std": 0.08815218508243561, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3035937547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "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": 797.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 521.6875, "completions/mean_terminated_length": 521.6875, "completions/min_length": 362.0, "completions/min_terminated_length": 362.0, "epoch": 0.2244, "grad_norm": 0.06368713080883026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12558746.0, "reward": 0.7687508463859558, "reward_std": 0.3376949429512024, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.4124999940395355, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7562500238418579, "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": 624.0, "completions/max_terminated_length": 624.0, "completions/mean_length": 484.9375, "completions/mean_terminated_length": 484.9375, "completions/min_length": 330.0, "completions/min_terminated_length": 330.0, "epoch": 0.2246, "grad_norm": 0.029171530157327652, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12570137.0, "reward": 0.7644537687301636, "reward_std": 0.14648909866809845, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5289062261581421, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6468750238418579, "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": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 686.4375, "completions/mean_terminated_length": 686.4375, "completions/min_length": 462.0, "completions/min_terminated_length": 462.0, "epoch": 0.2248, "grad_norm": 0.024660950526595116, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12585920.0, "reward": 0.5702352523803711, "reward_std": 0.17819823324680328, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2029687464237213, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8687499761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 439.0625, "completions/mean_terminated_length": 400.0666809082031, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.225, "grad_norm": 0.01970498263835907, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12597313.0, "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": 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": 448.0, "completions/max_terminated_length": 448.0, "completions/mean_length": 282.5625, "completions/mean_terminated_length": 282.5625, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.2252, "grad_norm": 0.0041848295368254185, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12605466.0, "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": 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": 792.0, "completions/max_terminated_length": 792.0, "completions/mean_length": 520.125, "completions/mean_terminated_length": 520.125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.2254, "grad_norm": 0.06608501076698303, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12617580.0, "reward": 0.7483607530593872, "reward_std": 0.39091259241104126, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.30921876430511475, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.909375011920929, "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": 635.0, "completions/max_terminated_length": 635.0, "completions/mean_length": 509.9375, "completions/mean_terminated_length": 509.9375, "completions/min_length": 399.0, "completions/min_terminated_length": 399.0, "epoch": 0.2256, "grad_norm": 0.10671895742416382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12629835.0, "reward": 1.0752885341644287, "reward_std": 0.4962671995162964, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5880749821662903, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9599999785423279, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 837.0, "completions/mean_length": 486.875, "completions/mean_terminated_length": 451.0666809082031, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 0.2258, "grad_norm": 0.0942825973033905, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12641689.0, "reward": 0.805391788482666, "reward_std": 0.5012853741645813, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3607812523841858, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.903124988079071, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 877.0, "completions/mean_length": 612.3125, "completions/mean_terminated_length": 584.86669921875, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.226, "grad_norm": 0.1103978306055069, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12655358.0, "reward": 0.943281888961792, "reward_std": 0.5593025088310242, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6990624666213989, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.6156250238418579, "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": 296.0, "completions/max_terminated_length": 296.0, "completions/mean_length": 216.6875, "completions/mean_terminated_length": 216.6875, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.2262, "grad_norm": 0.13942453265190125, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12662121.0, "reward": 0.7070823907852173, "reward_std": 0.3938066065311432, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2266625016927719, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 596.6875, "completions/mean_terminated_length": 568.2000122070312, "completions/min_length": 367.0, "completions/min_terminated_length": 367.0, "epoch": 0.2264, "grad_norm": 0.02485366351902485, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12675668.0, "reward": 0.5125542879104614, "reward_std": 0.1564820110797882, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.08760624378919601, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8881250023841858, "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": 695.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 413.0, "completions/mean_terminated_length": 413.0, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.2266, "grad_norm": 0.09385162591934204, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12686436.0, "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": 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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 717.8125, "completions/mean_terminated_length": 647.1538696289062, "completions/min_length": 465.0, "completions/min_terminated_length": 465.0, "epoch": 0.2268, "grad_norm": 0.027950512245297432, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12703057.0, "reward": 0.4410039484500885, "reward_std": 0.21921779215335846, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06950624287128448, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7768750190734863, "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": 629.0, "completions/max_terminated_length": 629.0, "completions/mean_length": 469.5, "completions/mean_terminated_length": 469.5, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 0.227, "grad_norm": 0.08876986056566238, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12715161.0, "reward": 0.8908482789993286, "reward_std": 0.48768168687820435, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4066937565803528, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9818750023841858, "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": 857.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 641.125, "completions/mean_terminated_length": 641.125, "completions/min_length": 447.0, "completions/min_terminated_length": 447.0, "epoch": 0.2272, "grad_norm": 0.050449032336473465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12729355.0, "reward": 0.7325946688652039, "reward_std": 0.3842404782772064, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2776874899864197, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 958.5625, "completions/mean_terminated_length": 893.125, "completions/min_length": 680.0, "completions/min_terminated_length": 680.0, "epoch": 0.2274, "grad_norm": 0.025376413017511368, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12749716.0, "reward": 0.2400786131620407, "reward_std": 0.28115114569664, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.042656250298023224, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.4375, "rewards/mean_confidence_reward": 0.4156249761581421, "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": 644.0, "completions/max_terminated_length": 644.0, "completions/mean_length": 429.625, "completions/mean_terminated_length": 429.625, "completions/min_length": 343.0, "completions/min_terminated_length": 343.0, "epoch": 0.2276, "grad_norm": 0.11928956210613251, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12760894.0, "reward": 0.8456259965896606, "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": 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": 634.0, "completions/max_terminated_length": 634.0, "completions/mean_length": 477.0625, "completions/mean_terminated_length": 477.0625, "completions/min_length": 329.0, "completions/min_terminated_length": 329.0, "epoch": 0.2278, "grad_norm": 0.04951570928096771, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12772591.0, "reward": 0.7221100330352783, "reward_std": 0.30192831158638, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.4442187547683716, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.715624988079071, "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": 0.0, "completions/max_length": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 449.625, "completions/mean_terminated_length": 449.625, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.228, "grad_norm": 0.08273564279079437, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12784217.0, "reward": 1.2570416927337646, "reward_std": 0.4340568780899048, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7640812397003174, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756250381469727, "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": 855.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 491.5, "completions/mean_terminated_length": 491.5, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 0.2282, "grad_norm": 0.09089326858520508, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12795873.0, "reward": 1.2616885900497437, "reward_std": 0.4251141846179962, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7733749747276306, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 657.0, "completions/max_terminated_length": 657.0, "completions/mean_length": 467.25, "completions/mean_terminated_length": 467.25, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.2284, "grad_norm": 0.09504160284996033, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12807893.0, "reward": 0.7541886568069458, "reward_std": 0.3968888521194458, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.25837498903274536, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9087499976158142, "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": 0.0, "completions/max_length": 642.0, "completions/max_terminated_length": 642.0, "completions/mean_length": 401.1875, "completions/mean_terminated_length": 401.1875, "completions/min_length": 348.0, "completions/min_terminated_length": 348.0, "epoch": 0.2286, "grad_norm": 0.11097181588411331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12818344.0, "reward": 0.9590573310852051, "reward_std": 0.49209892749786377, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.48061251640319824, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 0.0, "completions/max_length": 691.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 495.3125, "completions/mean_terminated_length": 495.3125, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.2288, "grad_norm": 0.11309295147657394, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12829853.0, "reward": 0.8042820692062378, "reward_std": 0.4187149405479431, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.48356249928474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7450000047683716, "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": 0.0, "completions/max_length": 543.0, "completions/max_terminated_length": 543.0, "completions/mean_length": 334.0, "completions/mean_terminated_length": 334.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 0.229, "grad_norm": 0.1255512833595276, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12839373.0, "reward": 0.9771103858947754, "reward_std": 0.47765469551086426, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5167187452316284, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.921875, "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": 0.5, "completions/max_length": 1024.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 871.0, "completions/mean_terminated_length": 718.0, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.2292, "grad_norm": 0.06877780705690384, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12857133.0, "reward": 0.4558597505092621, "reward_std": 0.5428943634033203, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2867187559604645, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.5, "rewards/mean_confidence_reward": 0.3843750059604645, "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": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 642.625, "completions/mean_terminated_length": 642.625, "completions/min_length": 528.0, "completions/min_terminated_length": 528.0, "epoch": 0.2294, "grad_norm": 0.06577331572771072, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12871943.0, "reward": 0.9526917338371277, "reward_std": 0.49854031205177307, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4678812325000763, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781249761581421, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 684.875, "completions/mean_terminated_length": 571.8333740234375, "completions/min_length": 416.0, "completions/min_terminated_length": 416.0, "epoch": 0.2296, "grad_norm": 0.1278039664030075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12886933.0, "reward": 0.7783443927764893, "reward_std": 0.6230109930038452, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43168750405311584, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.7012500166893005, "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": 669.0, "completions/max_terminated_length": 669.0, "completions/mean_length": 450.125, "completions/mean_terminated_length": 450.125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 0.2298, "grad_norm": 0.04045705497264862, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12898151.0, "reward": 0.6045072078704834, "reward_std": 0.23866797983646393, "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": 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": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 306.0625, "completions/mean_terminated_length": 306.0625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 0.23, "grad_norm": 0.10488247126340866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12907016.0, "reward": 1.3208446502685547, "reward_std": 0.3830661177635193, "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": 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": 724.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 471.75, "completions/mean_terminated_length": 471.75, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.2302, "grad_norm": 0.14916115999221802, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 12918660.0, "reward": 1.1031792163848877, "reward_std": 0.5431605577468872, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6438562870025635, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9743750095367432, "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": 0.0, "completions/max_length": 610.0, "completions/max_terminated_length": 610.0, "completions/mean_length": 416.25, "completions/mean_terminated_length": 416.25, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.2304, "grad_norm": 0.006725590210407972, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12929128.0, "reward": 0.5329511761665344, "reward_std": 0.02157808467745781, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06589999794960022, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500023841858, "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": 407.0, "completions/max_terminated_length": 407.0, "completions/mean_length": 293.25, "completions/mean_terminated_length": 293.25, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.2306, "grad_norm": 0.11145688593387604, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12937388.0, "reward": 0.7796604037284851, "reward_std": 0.4292029142379761, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.30931875109672546, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 395.0, "completions/max_terminated_length": 395.0, "completions/mean_length": 276.6875, "completions/mean_terminated_length": 276.6875, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.2308, "grad_norm": 0.17706672847270966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12945895.0, "reward": 1.1970763206481934, "reward_std": 0.46379709243774414, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7066500186920166, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 444.0, "completions/max_terminated_length": 444.0, "completions/mean_length": 343.875, "completions/mean_terminated_length": 343.875, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.231, "grad_norm": 0.13008731603622437, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12954853.0, "reward": 1.0152323246002197, "reward_std": 0.49963855743408203, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5304625034332275, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587499499320984, "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": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 752.875, "completions/mean_terminated_length": 752.875, "completions/min_length": 422.0, "completions/min_terminated_length": 422.0, "epoch": 0.2312, "grad_norm": 0.04389696195721626, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12971315.0, "reward": 0.7897478342056274, "reward_std": 0.42353978753089905, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.32949376106262207, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "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": 0.0, "completions/max_length": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 623.0, "completions/mean_terminated_length": 623.0, "completions/min_length": 503.0, "completions/min_terminated_length": 503.0, "epoch": 0.2314, "grad_norm": 0.0657569020986557, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12985107.0, "reward": 1.1403697729110718, "reward_std": 0.4785560369491577, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6557375192642212, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9624999761581421, "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": 0.0, "completions/max_length": 354.0, "completions/max_terminated_length": 354.0, "completions/mean_length": 227.625, "completions/mean_terminated_length": 227.625, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "epoch": 0.2316, "grad_norm": 0.22706294059753418, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 12992285.0, "reward": 1.2547292709350586, "reward_std": 0.438401997089386, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7594562768936157, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9831250309944153, "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": 478.0, "completions/max_terminated_length": 478.0, "completions/mean_length": 371.5, "completions/mean_terminated_length": 371.5, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.2318, "grad_norm": 0.09394457936286926, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13002021.0, "reward": 1.0797321796417236, "reward_std": 0.49125441908836365, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5969625115394592, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9612500071525574, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 777.375, "completions/mean_terminated_length": 760.933349609375, "completions/min_length": 555.0, "completions/min_terminated_length": 555.0, "epoch": 0.232, "grad_norm": 0.030910708010196686, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13018763.0, "reward": 0.703047513961792, "reward_std": 0.2796434760093689, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4685937464237213, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.565625011920929, "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": 754.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 483.6875, "completions/mean_terminated_length": 483.6875, "completions/min_length": 400.0, "completions/min_terminated_length": 400.0, "epoch": 0.2322, "grad_norm": 0.0998321920633316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13031510.0, "reward": 1.0174229145050049, "reward_std": 0.496869295835495, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5348437428474426, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "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": 615.0, "completions/max_terminated_length": 615.0, "completions/mean_length": 471.9375, "completions/mean_terminated_length": 471.9375, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.2324, "grad_norm": 0.09428563714027405, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13043173.0, "reward": 1.1921823024749756, "reward_std": 0.4711289405822754, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.696862518787384, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 767.3125, "completions/mean_terminated_length": 681.75, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.2326, "grad_norm": 0.05545981228351593, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13059786.0, "reward": 0.5630474090576172, "reward_std": 0.4273688793182373, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.4385937452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.33125001192092896, "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": 915.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 604.75, "completions/mean_terminated_length": 604.75, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 0.2328, "grad_norm": 0.053274501115083694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13073382.0, "reward": 1.3489854335784912, "reward_std": 0.32823508977890015, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8854687213897705, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 499.9375, "completions/mean_terminated_length": 465.0000305175781, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.233, "grad_norm": 0.10589489340782166, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13085397.0, "reward": 1.115066409111023, "reward_std": 0.5209187269210815, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6676312685012817, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8518750071525574, "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": 866.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 583.25, "completions/mean_terminated_length": 583.25, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.2332, "grad_norm": 0.029747826978564262, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13099529.0, "reward": 0.8013163208961487, "reward_std": 0.19594421982765198, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6026312112808228, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.5081250071525574, "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": 0.0, "completions/max_length": 525.0, "completions/max_terminated_length": 525.0, "completions/mean_length": 346.5625, "completions/mean_terminated_length": 346.5625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 0.2334, "grad_norm": 0.11440374702215195, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13108930.0, "reward": 1.1395978927612305, "reward_std": 0.4431816041469574, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7166937589645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8331249952316284, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 718.875, "completions/mean_terminated_length": 675.2857666015625, "completions/min_length": 456.0, "completions/min_terminated_length": 456.0, "epoch": 0.2336, "grad_norm": 0.05291778966784477, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13124192.0, "reward": 0.618672788143158, "reward_std": 0.35770371556282043, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23734375834465027, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.640625, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 807.25, "completions/mean_terminated_length": 735.0, "completions/min_length": 597.0, "completions/min_terminated_length": 597.0, "epoch": 0.2338, "grad_norm": 0.10803133994340897, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13141524.0, "reward": 0.8062131404876709, "reward_std": 0.6745358109474182, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4874249994754791, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.6875, "rewards/mean_confidence_reward": 0.6399999856948853, "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": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 533.0, "completions/mean_terminated_length": 533.0, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 0.234, "grad_norm": 0.08861099928617477, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13153700.0, "reward": 1.2006856203079224, "reward_std": 0.45802974700927734, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7138687372207642, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9756249785423279, "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": 1017.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 585.3125, "completions/mean_terminated_length": 585.3125, "completions/min_length": 317.0, "completions/min_terminated_length": 317.0, "epoch": 0.2342, "grad_norm": 0.09404972195625305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13167353.0, "reward": 0.9799102544784546, "reward_std": 0.5024062395095825, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5848187208175659, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8924999833106995, "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": 0.0, "completions/max_length": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 499.5625, "completions/mean_terminated_length": 499.5625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2344, "grad_norm": 0.1037411019206047, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13179586.0, "reward": 1.289610505104065, "reward_std": 0.3878503739833832, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8292187452316284, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968749642372131, "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": 562.0, "completions/max_terminated_length": 562.0, "completions/mean_length": 299.5625, "completions/mean_terminated_length": 299.5625, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.2346, "grad_norm": 0.058081842958927155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13190171.0, "reward": 0.852031946182251, "reward_std": 0.16290809214115143, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.7040625214576721, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4375, "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": 507.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 375.3125, "completions/mean_terminated_length": 375.3125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.2348, "grad_norm": 0.06493605673313141, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13200560.0, "reward": 0.6166415214538574, "reward_std": 0.2365012764930725, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.17078125476837158, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.940625011920929, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 507.0, "completions/mean_length": 456.75, "completions/mean_terminated_length": 375.71429443359375, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 0.235, "grad_norm": 0.1300417184829712, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13211756.0, "reward": 0.72304767370224, "reward_std": 0.4655528664588928, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.38359373807907104, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7093749642372131, "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": 503.0, "completions/max_terminated_length": 503.0, "completions/mean_length": 383.4375, "completions/mean_terminated_length": 383.4375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.2352, "grad_norm": 0.07965535670518875, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13221875.0, "reward": 0.6656915545463562, "reward_std": 0.3252728581428528, "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": 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": 455.0, "completions/max_terminated_length": 455.0, "completions/mean_length": 310.3125, "completions/mean_terminated_length": 310.3125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 0.2354, "grad_norm": 0.17276175320148468, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13231048.0, "reward": 1.0190259218215942, "reward_std": 0.49589434266090393, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5380500555038452, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 662.9375, "completions/mean_terminated_length": 638.86669921875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2356, "grad_norm": 0.052446600049734116, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13245351.0, "reward": 0.6747633218765259, "reward_std": 0.36701929569244385, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.28702500462532043, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8399999737739563, "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": 948.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 751.25, "completions/mean_terminated_length": 751.25, "completions/min_length": 501.0, "completions/min_terminated_length": 501.0, "epoch": 0.2358, "grad_norm": 0.09101832658052444, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13261563.0, "reward": 1.1079823970794678, "reward_std": 0.5379907488822937, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6534624695777893, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9399999976158142, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 765.0625, "completions/mean_terminated_length": 705.3077392578125, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 0.236, "grad_norm": 0.04370682314038277, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13278124.0, "reward": 0.5701567530632019, "reward_std": 0.3705005645751953, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3903124928474426, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.5062500238418579, "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": 0.0, "completions/max_length": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 435.8125, "completions/mean_terminated_length": 435.8125, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 0.2362, "grad_norm": 0.06975360214710236, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13288793.0, "reward": 0.5941355228424072, "reward_std": 0.24216853082180023, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1257687509059906, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9643750190734863, "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": 462.0, "completions/max_terminated_length": 462.0, "completions/mean_length": 367.125, "completions/mean_terminated_length": 367.125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 0.2364, "grad_norm": 0.00014966427988838404, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13298891.0, "reward": 1.4993321895599365, "reward_std": 0.0006022243178449571, "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": 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": 447.0, "completions/max_terminated_length": 447.0, "completions/mean_length": 326.875, "completions/mean_terminated_length": 326.875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.2366, "grad_norm": 0.11662263423204422, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13308505.0, "reward": 1.0120729207992554, "reward_std": 0.5037750601768494, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5241437554359436, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9781250357627869, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 680.0, "completions/mean_terminated_length": 600.6154174804688, "completions/min_length": 427.0, "completions/min_terminated_length": 427.0, "epoch": 0.2368, "grad_norm": 0.09757334738969803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13323577.0, "reward": 0.9379695653915405, "reward_std": 0.6218911409378052, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5634375214576721, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7437499761581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 589.25, "completions/mean_terminated_length": 560.2667236328125, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.237, "grad_norm": 0.08804164081811905, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13337421.0, "reward": 0.8612010478973389, "reward_std": 0.5270626544952393, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4099000096321106, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9075000286102295, "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": 463.0, "completions/max_terminated_length": 463.0, "completions/mean_length": 277.5625, "completions/mean_terminated_length": 277.5625, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.2372, "grad_norm": 0.13828310370445251, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13345606.0, "reward": 1.1498823165893555, "reward_std": 0.469413697719574, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.674762487411499, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 594.0, "completions/max_terminated_length": 594.0, "completions/mean_length": 516.5, "completions/mean_terminated_length": 516.5, "completions/min_length": 363.0, "completions/min_terminated_length": 363.0, "epoch": 0.2374, "grad_norm": 0.07612995058298111, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13357982.0, "reward": 0.9606226086616516, "reward_std": 0.4251859486103058, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6712437868118286, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6899999976158142, "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": 540.0, "completions/max_terminated_length": 540.0, "completions/mean_length": 403.875, "completions/mean_terminated_length": 403.875, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 0.2376, "grad_norm": 0.07777262479066849, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13368268.0, "reward": 0.6015385389328003, "reward_std": 0.2400481253862381, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14057499170303345, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 887.0, "completions/mean_length": 679.875, "completions/mean_terminated_length": 656.933349609375, "completions/min_length": 414.0, "completions/min_terminated_length": 414.0, "epoch": 0.2378, "grad_norm": 0.04955678805708885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13383162.0, "reward": 0.6429698467254639, "reward_std": 0.36659711599349976, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.22343748807907104, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8812499642372131, "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": 298.0, "completions/max_terminated_length": 298.0, "completions/mean_length": 229.4375, "completions/mean_terminated_length": 229.4375, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.238, "grad_norm": 0.19181694090366364, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13390465.0, "reward": 1.250991940498352, "reward_std": 0.4448452591896057, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.751981258392334, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9856250286102295, "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": 659.0, "completions/max_terminated_length": 659.0, "completions/mean_length": 483.3125, "completions/mean_terminated_length": 483.3125, "completions/min_length": 401.0, "completions/min_terminated_length": 401.0, "epoch": 0.2382, "grad_norm": 0.09400341659784317, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13402582.0, "reward": 0.7065635919570923, "reward_std": 0.33550745248794556, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.2881249785423279, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.875, "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": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 475.875, "completions/mean_terminated_length": 475.875, "completions/min_length": 361.0, "completions/min_terminated_length": 361.0, "epoch": 0.2384, "grad_norm": 0.0899253562092781, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13414164.0, "reward": 1.1407572031021118, "reward_std": 0.4775138199329376, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6565124988555908, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9537500143051147, "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": 0.0, "completions/max_length": 918.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 493.0, "completions/mean_terminated_length": 493.0, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.2386, "grad_norm": 0.08846773952245712, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13426004.0, "reward": 0.7003884315490723, "reward_std": 0.41322779655456543, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2757750153541565, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8525000214576721, "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": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 391.8125, "completions/mean_terminated_length": 391.8125, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 0.2388, "grad_norm": 0.02245614118874073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13436449.0, "reward": 0.5644447803497314, "reward_std": 0.08850265294313431, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.12888750433921814, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262499809265137, "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": 988.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 718.875, "completions/mean_terminated_length": 718.875, "completions/min_length": 444.0, "completions/min_terminated_length": 444.0, "epoch": 0.239, "grad_norm": 0.06897792220115662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13452063.0, "reward": 0.8547540307044983, "reward_std": 0.4523892104625702, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45950624346733093, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7581250071525574, "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": 534.0, "completions/max_terminated_length": 534.0, "completions/mean_length": 320.0, "completions/mean_terminated_length": 320.0, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.2392, "grad_norm": 0.053717274218797684, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13461583.0, "reward": 1.4327919483184814, "reward_std": 0.24689167737960815, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.928081214427948, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9593750238418579, "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": 0.0, "completions/max_length": 933.0, "completions/max_terminated_length": 933.0, "completions/mean_length": 690.875, "completions/mean_terminated_length": 690.875, "completions/min_length": 438.0, "completions/min_terminated_length": 438.0, "epoch": 0.2394, "grad_norm": 0.03628357872366905, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13479309.0, "reward": 0.7834380865097046, "reward_std": 0.20856082439422607, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.5668749809265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.543749988079071, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 723.1875, "completions/mean_terminated_length": 680.2142944335938, "completions/min_length": 467.0, "completions/min_terminated_length": 467.0, "epoch": 0.2396, "grad_norm": 0.036116596311330795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13494784.0, "reward": 0.5392853617668152, "reward_std": 0.3177366554737091, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.14106875658035278, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8331249952316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 482.0, "completions/mean_length": 380.3125, "completions/mean_terminated_length": 337.4000244140625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 0.2398, "grad_norm": 0.11103416234254837, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13504485.0, "reward": 0.7510948777198792, "reward_std": 0.46781566739082336, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31468749046325684, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.893750011920929, "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": 548.0, "completions/max_terminated_length": 548.0, "completions/mean_length": 377.0, "completions/mean_terminated_length": 377.0, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.24, "grad_norm": 0.10713478177785873, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13514933.0, "reward": 0.7822355031967163, "reward_std": 0.4279831051826477, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.31446874141693115, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9631249904632568, "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": 619.0, "completions/max_terminated_length": 619.0, "completions/mean_length": 384.75, "completions/mean_terminated_length": 384.75, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 0.2402, "grad_norm": 0.06951484829187393, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13524977.0, "reward": 0.5202353000640869, "reward_std": 0.1396171897649765, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.1029687449336052, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 652.0, "completions/mean_length": 374.125, "completions/mean_terminated_length": 330.8000183105469, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 0.2404, "grad_norm": 0.1021246686577797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13535139.0, "reward": 0.7517977952957153, "reward_std": 0.46590372920036316, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3160937428474426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.890625, "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": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 314.875, "completions/mean_terminated_length": 314.875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 0.2406, "grad_norm": 0.04372718557715416, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13544081.0, "reward": 0.6193634271621704, "reward_std": 0.15151338279247284, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.2387249916791916, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8362500071525574, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 590.8125, "completions/mean_terminated_length": 561.933349609375, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.2408, "grad_norm": 0.10129205137491226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13558126.0, "reward": 0.7693756818771362, "reward_std": 0.46016669273376465, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.6012499928474426, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.40937501192092896, "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": 694.0, "completions/max_terminated_length": 694.0, "completions/mean_length": 471.3125, "completions/mean_terminated_length": 471.3125, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.241, "grad_norm": 0.05542408674955368, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13571155.0, "reward": 0.8610943555831909, "reward_std": 0.23841363191604614, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.6596875190734863, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.550000011920929, "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": 0.4375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 798.0, "completions/mean_terminated_length": 622.2222290039062, "completions/min_length": 434.0, "completions/min_terminated_length": 434.0, "epoch": 0.2412, "grad_norm": 0.06870413571596146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13587843.0, "reward": 0.35586017370224, "reward_std": 0.402043879032135, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.08671875298023224, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.5625, "rewards/mean_confidence_reward": 0.546875, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 691.8125, "completions/mean_terminated_length": 669.6666870117188, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 0.2414, "grad_norm": 0.024372467771172523, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13602864.0, "reward": 0.5464041233062744, "reward_std": 0.186960369348526, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15530624985694885, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8368749618530273, "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": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 419.625, "completions/mean_terminated_length": 419.625, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.2416, "grad_norm": 0.10479568690061569, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13614170.0, "reward": 1.1992073059082031, "reward_std": 0.45984670519828796, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7109124660491943, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9662500619888306, "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": 349.0, "completions/max_terminated_length": 349.0, "completions/mean_length": 264.6875, "completions/mean_terminated_length": 264.6875, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.2418, "grad_norm": 0.07960930466651917, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13622469.0, "reward": 0.6253134608268738, "reward_std": 0.23457297682762146, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.18812499940395355, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9312499761581421, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 665.5625, "completions/mean_terminated_length": 614.357177734375, "completions/min_length": 406.0, "completions/min_terminated_length": 406.0, "epoch": 0.242, "grad_norm": 0.08463867008686066, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13637966.0, "reward": 0.8959883451461792, "reward_std": 0.5771682858467102, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4794750213623047, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8374999761581421, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 504.625, "completions/mean_terminated_length": 430.4285888671875, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.2422, "grad_norm": 0.1069110631942749, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13650104.0, "reward": 0.9524571895599365, "reward_std": 0.590340256690979, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.529912531375885, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8450000286102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 500.1875, "completions/mean_terminated_length": 465.2666931152344, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.2424, "grad_norm": 0.08636517077684402, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13662443.0, "reward": 0.5673447847366333, "reward_std": 0.2828725278377533, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.13468749821186066, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8937499523162842, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 582.0625, "completions/mean_terminated_length": 552.6000366210938, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 0.2426, "grad_norm": 0.031395506113767624, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13676156.0, "reward": 0.46765440702438354, "reward_std": 0.1879749298095703, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.06030625104904175, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 442.1875, "completions/mean_terminated_length": 403.4000244140625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.2428, "grad_norm": 0.10090778023004532, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13687055.0, "reward": 0.614532470703125, "reward_std": 0.3218234181404114, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2290624976158142, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7937500476837158, "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": 554.0, "completions/max_terminated_length": 554.0, "completions/mean_length": 391.0625, "completions/mean_terminated_length": 391.0625, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.243, "grad_norm": 0.0910576656460762, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13697904.0, "reward": 0.7297638654708862, "reward_std": 0.3892582952976227, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2720249891281128, "rewards/confidence_one_or_zero": 0.4375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "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": 501.0, "completions/max_terminated_length": 501.0, "completions/mean_length": 358.5625, "completions/mean_terminated_length": 358.5625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.2432, "grad_norm": 0.12415794283151627, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13707193.0, "reward": 1.2588258981704712, "reward_std": 0.42927858233451843, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7676500082015991, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 429.0, "completions/max_terminated_length": 429.0, "completions/mean_length": 351.25, "completions/mean_terminated_length": 351.25, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.2434, "grad_norm": 0.06845097243785858, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13716925.0, "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": 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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 512.75, "completions/mean_terminated_length": 478.66668701171875, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.2436, "grad_norm": 0.09738007932901382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13728889.0, "reward": 0.5989820957183838, "reward_std": 0.39742210507392883, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.19796249270439148, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8899999856948853, "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": 469.0, "completions/max_terminated_length": 469.0, "completions/mean_length": 334.375, "completions/mean_terminated_length": 334.375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.2438, "grad_norm": 0.12307647615671158, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13737759.0, "reward": 1.2249010801315308, "reward_std": 0.505244791507721, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7623000144958496, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9674999713897705, "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": 531.0, "completions/max_terminated_length": 531.0, "completions/mean_length": 358.1875, "completions/mean_terminated_length": 358.1875, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.244, "grad_norm": 0.10781879723072052, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13747842.0, "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": 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": 218.0, "completions/max_terminated_length": 218.0, "completions/mean_length": 169.0625, "completions/mean_terminated_length": 169.0625, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.2442, "grad_norm": 0.3270707130432129, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13754003.0, "reward": 1.0120823383331299, "reward_std": 0.5037587881088257, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5241625308990479, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9787499904632568, "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": 458.0, "completions/max_terminated_length": 458.0, "completions/mean_length": 388.1875, "completions/mean_terminated_length": 388.1875, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 0.2444, "grad_norm": 0.07742283493280411, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13764230.0, "reward": 1.38021981716156, "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": 0.9575000405311584, "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": 306.0, "completions/max_terminated_length": 306.0, "completions/mean_length": 255.4375, "completions/mean_terminated_length": 255.4375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.2446, "grad_norm": 0.1407889425754547, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13772285.0, "reward": 1.0733041763305664, "reward_std": 0.4994734525680542, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5841062664985657, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9768750071525574, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 827.125, "completions/mean_terminated_length": 781.6923217773438, "completions/min_length": 499.0, "completions/min_terminated_length": 499.0, "epoch": 0.2448, "grad_norm": 0.0673135370016098, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13790255.0, "reward": 0.6492072343826294, "reward_std": 0.4950508177280426, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2984125018119812, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7237499952316284, "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": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 477.0625, "completions/mean_terminated_length": 477.0625, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 0.245, "grad_norm": 0.1580273061990738, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13801584.0, "reward": 1.07834792137146, "reward_std": 0.492647647857666, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5941938161849976, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9581249952316284, "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": 474.0, "completions/max_terminated_length": 474.0, "completions/mean_length": 342.125, "completions/mean_terminated_length": 342.125, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 0.2452, "grad_norm": 0.12034734338521957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13810978.0, "reward": 1.3207696676254272, "reward_std": 0.3830287754535675, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8290375471115112, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 359.0, "completions/max_terminated_length": 359.0, "completions/mean_length": 274.0, "completions/mean_terminated_length": 274.0, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.2454, "grad_norm": 0.059339847415685654, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13819090.0, "reward": 0.8342976570129395, "reward_std": 0.17905402183532715, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.6685937643051147, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.4468749761581421, "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": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 411.8125, "completions/mean_terminated_length": 411.8125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.2456, "grad_norm": 0.09556882083415985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13829199.0, "reward": 1.3209354877471924, "reward_std": 0.38311120867729187, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8293687105178833, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9618749618530273, "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": 0.0, "completions/max_length": 470.0, "completions/max_terminated_length": 470.0, "completions/mean_length": 364.125, "completions/mean_terminated_length": 364.125, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.2458, "grad_norm": 0.10622182488441467, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13838881.0, "reward": 1.314732551574707, "reward_std": 0.39766016602516174, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.816962480545044, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9850000143051147, "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": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 528.75, "completions/mean_terminated_length": 528.75, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 0.246, "grad_norm": 0.058266282081604004, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13851021.0, "reward": 0.9299882054328918, "reward_std": 0.3646262586116791, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.6099749803543091, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7487499713897705, "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": 473.0, "completions/max_terminated_length": 473.0, "completions/mean_length": 339.5, "completions/mean_terminated_length": 339.5, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.2462, "grad_norm": 0.10388662666082382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13860389.0, "reward": 1.110079050064087, "reward_std": 0.4363798499107361, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6576562523841858, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.828125, "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": 456.0, "completions/max_terminated_length": 456.0, "completions/mean_length": 360.375, "completions/mean_terminated_length": 360.375, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.2464, "grad_norm": 0.11467919498682022, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13869851.0, "reward": 1.0792416334152222, "reward_std": 0.4921223521232605, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.595981240272522, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.965624988079071, "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": 0.0, "completions/max_length": 343.0, "completions/max_terminated_length": 343.0, "completions/mean_length": 248.8125, "completions/mean_terminated_length": 248.8125, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.2466, "grad_norm": 0.17980483174324036, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13877496.0, "reward": 0.8920698165893555, "reward_std": 0.48646309971809387, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.40913742780685425, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.976187527179718, "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": 779.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 516.4375, "completions/mean_terminated_length": 516.4375, "completions/min_length": 308.0, "completions/min_terminated_length": 308.0, "epoch": 0.2468, "grad_norm": 0.02229210175573826, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13889551.0, "reward": 0.5760167241096497, "reward_std": 0.12002433836460114, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.15203124284744263, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.890625, "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": 0.0, "completions/max_length": 687.0, "completions/max_terminated_length": 687.0, "completions/mean_length": 454.0625, "completions/mean_terminated_length": 454.0625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.247, "grad_norm": 0.12318920344114304, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13900656.0, "reward": 0.9670977592468262, "reward_std": 0.4840315580368042, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4966937303543091, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "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": 726.0, "completions/max_terminated_length": 726.0, "completions/mean_length": 528.4375, "completions/mean_terminated_length": 528.4375, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.2472, "grad_norm": 0.07911141961812973, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13912935.0, "reward": 1.206632137298584, "reward_std": 0.40516197681427, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7882624864578247, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8637499809265137, "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": 0.25, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 720.9375, "completions/mean_terminated_length": 619.9166870117188, "completions/min_length": 385.0, "completions/min_terminated_length": 385.0, "epoch": 0.2474, "grad_norm": 0.07581453770399094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13928262.0, "reward": 0.4895320534706116, "reward_std": 0.37457630038261414, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.1665624976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.75, "rewards/mean_confidence_reward": 0.6937500238418579, "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": 831.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 482.8125, "completions/mean_terminated_length": 482.8125, "completions/min_length": 359.0, "completions/min_terminated_length": 359.0, "epoch": 0.2476, "grad_norm": 0.07457055151462555, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13939651.0, "reward": 1.2600510120391846, "reward_std": 0.42768311500549316, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.7700999975204468, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9562499523162842, "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": 387.0, "completions/max_terminated_length": 387.0, "completions/mean_length": 257.6875, "completions/mean_terminated_length": 257.6875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.2478, "grad_norm": 0.00017380084318574518, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13948718.0, "reward": 1.4991071224212646, "reward_std": 0.0005474342033267021, "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": 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": 0.0, "completions/max_length": 506.0, "completions/max_terminated_length": 506.0, "completions/mean_length": 384.9375, "completions/mean_terminated_length": 384.9375, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.248, "grad_norm": 0.1037178784608841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13958573.0, "reward": 0.8432666063308716, "reward_std": 0.4566669464111328, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3740312457084656, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "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": 378.0, "completions/max_terminated_length": 378.0, "completions/mean_length": 250.125, "completions/mean_terminated_length": 250.125, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.2482, "grad_norm": 0.06997527182102203, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13966175.0, "reward": 1.4354572296142578, "reward_std": 0.2367425411939621, "rewards/accuracy_reward": 0.9375, "rewards/brier_reward": 0.933412492275238, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9350000619888306, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 527.3125, "completions/mean_terminated_length": 494.20001220703125, "completions/min_length": 377.0, "completions/min_terminated_length": 377.0, "epoch": 0.2484, "grad_norm": 0.07573731243610382, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 13978340.0, "reward": 1.2694289684295654, "reward_std": 0.4519370198249817, "rewards/accuracy_reward": 0.75, "rewards/brier_reward": 0.8513562679290771, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8318749666213989, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 650.75, "completions/mean_terminated_length": 597.4285888671875, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 0.2486, "grad_norm": 0.10215246677398682, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 13992672.0, "reward": 0.9498447179794312, "reward_std": 0.5837514996528625, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5246875286102295, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 749.875, "completions/mean_terminated_length": 731.6000366210938, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 0.2488, "grad_norm": 0.079086072742939, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14008702.0, "reward": 0.9179699420928955, "reward_std": 0.54463130235672, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4609375, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.90625, "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": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 494.0, "completions/mean_terminated_length": 494.0, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 0.249, "grad_norm": 0.10483299940824509, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14020366.0, "reward": 0.8144416809082031, "reward_std": 0.49618902802467346, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37888121604919434, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9549999833106995, "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": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 632.3125, "completions/mean_terminated_length": 632.3125, "completions/min_length": 353.0, "completions/min_terminated_length": 353.0, "epoch": 0.2492, "grad_norm": 0.027201438322663307, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14034723.0, "reward": 0.6002728343009949, "reward_std": 0.24000659584999084, "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": 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": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 544.125, "completions/mean_terminated_length": 544.125, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.2494, "grad_norm": 0.06638963520526886, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14047269.0, "reward": 1.140075922012329, "reward_std": 0.3634122610092163, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.7176499962806702, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6650000214576721, "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": 234.0, "completions/max_terminated_length": 234.0, "completions/mean_length": 170.0, "completions/mean_terminated_length": 170.0, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.2496, "grad_norm": 0.20558544993400574, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14053653.0, "reward": 0.7555168867111206, "reward_std": 0.4437478184700012, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.26103127002716064, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9906250238418579, "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": 665.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 548.875, "completions/mean_terminated_length": 548.875, "completions/min_length": 491.0, "completions/min_terminated_length": 491.0, "epoch": 0.2498, "grad_norm": 0.06343350559473038, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14067251.0, "reward": 1.3808262348175049, "reward_std": 0.3248092830181122, "rewards/accuracy_reward": 0.875, "rewards/brier_reward": 0.886650025844574, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9800000190734863, "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": 553.0, "completions/max_terminated_length": 553.0, "completions/mean_length": 464.4375, "completions/mean_terminated_length": 464.4375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 0.25, "grad_norm": 0.1282631903886795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14078858.0, "reward": 1.0249228477478027, "reward_std": 0.4516419470310211, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6123437881469727, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8468749523162842, "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": 996.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 462.21875, "completions/mean_terminated_length": 462.21875, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 1.0008, "grad_norm": 0.036834392696619034, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14125864.0, "reward": 1.1012134552001953, "reward_std": 0.27780067920684814, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7024250030517578, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8143749833106995, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 481.1875, "completions/mean_terminated_length": 445.0000305175781, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 1.0016, "grad_norm": 0.053847167640924454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14172404.0, "reward": 0.9389370083808899, "reward_std": 0.4904528856277466, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5184968709945679, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8490625023841858, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 489.5625, "completions/mean_terminated_length": 481.07940673828125, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.0024, "grad_norm": 0.04461923614144325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14219464.0, "reward": 1.0627049207687378, "reward_std": 0.40640193223953247, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6566578149795532, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8245311975479126, "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": 967.0, "completions/max_terminated_length": 967.0, "completions/mean_length": 426.140625, "completions/mean_terminated_length": 426.140625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.0032, "grad_norm": 0.0452657975256443, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14263665.0, "reward": 0.9729752540588379, "reward_std": 0.32396644353866577, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5240734219551086, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9189062714576721, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 518.4375, "completions/mean_terminated_length": 493.57373046875, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.004, "grad_norm": 0.04473516345024109, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14315805.0, "reward": 0.7580001950263977, "reward_std": 0.38249820470809937, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3909984230995178, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7676562666893005, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 499.3125, "completions/mean_terminated_length": 490.9841613769531, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.0048, "grad_norm": 0.04115463048219681, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14364161.0, "reward": 0.7790564894676208, "reward_std": 0.3498268127441406, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.37061095237731934, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8890624642372131, "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": 0.0, "completions/max_length": 830.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 408.875, "completions/mean_terminated_length": 408.875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.0056, "grad_norm": 0.10948389023542404, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14405561.0, "reward": 1.0384931564331055, "reward_std": 0.4138428568840027, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6394844055175781, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.836718738079071, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 533.5625, "completions/mean_terminated_length": 492.0, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 1.0064, "grad_norm": 0.040918655693531036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14455725.0, "reward": 0.7647461295127869, "reward_std": 0.41038447618484497, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.4044906497001648, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7451562881469727, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 563.28125, "completions/mean_terminated_length": 555.96826171875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.0072, "grad_norm": 0.023334847763180733, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14506575.0, "reward": 0.6278424263000488, "reward_std": 0.22284090518951416, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2088078111410141, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8879687786102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 541.203125, "completions/mean_terminated_length": 533.5397338867188, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.008, "grad_norm": 0.026954999193549156, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14557196.0, "reward": 0.7949768304824829, "reward_std": 0.2992532253265381, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3555765748023987, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8970311880111694, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 539.046875, "completions/mean_terminated_length": 523.4031982421875, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.0088, "grad_norm": 0.03950145095586777, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14610255.0, "reward": 0.8917500376701355, "reward_std": 0.37245267629623413, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.5647484064102173, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7067188024520874, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 704.0, "completions/mean_length": 434.125, "completions/mean_terminated_length": 424.7619323730469, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 1.0096, "grad_norm": 0.028677426278591156, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14654119.0, "reward": 0.6834642291069031, "reward_std": 0.2697402834892273, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.27317655086517334, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8732812404632568, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 538.21875, "completions/mean_terminated_length": 522.5484008789062, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 1.0104, "grad_norm": 0.037215009331703186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14705717.0, "reward": 0.9335563778877258, "reward_std": 0.41226431727409363, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5389859080314636, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8235937356948853, "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": 1014.0, "completions/mean_length": 475.96875, "completions/mean_terminated_length": 439.433349609375, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.0112, "grad_norm": 0.06489291042089462, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14751475.0, "reward": 0.8647931814193726, "reward_std": 0.4574092626571655, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44833439588546753, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8606249690055847, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 441.609375, "completions/mean_terminated_length": 432.3651123046875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 1.012, "grad_norm": 0.04384024813771248, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14796538.0, "reward": 0.9649009704589844, "reward_std": 0.4113089442253113, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5860499739646912, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8081250190734863, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 434.71875, "completions/mean_terminated_length": 384.7796630859375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.0128, "grad_norm": 0.02997320331633091, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14840392.0, "reward": 1.0186681747436523, "reward_std": 0.2683437764644623, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5998343825340271, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8553124666213989, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 553.890625, "completions/mean_terminated_length": 514.0508422851562, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.0136, "grad_norm": 0.06678498536348343, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14891841.0, "reward": 0.726145327091217, "reward_std": 0.4239511489868164, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.43666407465934753, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6951562166213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 537.0, "completions/mean_length": 448.0625, "completions/mean_terminated_length": 429.4838562011719, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.0144, "grad_norm": 0.03338224068284035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 14936389.0, "reward": 1.0849337577819824, "reward_std": 0.28520968556404114, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.622990608215332, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9371875524520874, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 453.515625, "completions/mean_terminated_length": 444.4603576660156, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.0152, "grad_norm": 0.04973598197102547, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 14983062.0, "reward": 0.9418392181396484, "reward_std": 0.3798169493675232, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47742655873298645, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9437500238418579, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 556.78125, "completions/mean_terminated_length": 517.1864624023438, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.016, "grad_norm": 0.03141000494360924, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15037896.0, "reward": 0.6854617595672607, "reward_std": 0.30702435970306396, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.37092188000679016, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7000000476837158, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 510.15625, "completions/mean_terminated_length": 436.7500305175781, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 1.0168, "grad_norm": 0.03919941186904907, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15088034.0, "reward": 0.6846288442611694, "reward_std": 0.34756430983543396, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.40050622820854187, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6146875023841858, "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": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 403.15625, "completions/mean_terminated_length": 403.15625, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 1.0176, "grad_norm": 0.08025211840867996, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15129724.0, "reward": 0.8620947599411011, "reward_std": 0.3862864375114441, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4116874933242798, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9068750143051147, "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": 0.0, "completions/max_length": 666.0, "completions/max_terminated_length": 666.0, "completions/mean_length": 454.609375, "completions/mean_terminated_length": 454.609375, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 1.0184, "grad_norm": 0.05028204247355461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15175651.0, "reward": 1.107682228088379, "reward_std": 0.38062167167663574, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6216124892234802, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9496874809265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 800.0, "completions/mean_length": 510.921875, "completions/mean_terminated_length": 502.7778015136719, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 1.0192, "grad_norm": 0.040014248341321945, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15224574.0, "reward": 0.9734696745872498, "reward_std": 0.39346885681152344, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5563125014305115, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8506250381469727, "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": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 313.5, "completions/mean_terminated_length": 313.5, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "epoch": 1.02, "grad_norm": 0.05014047771692276, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15260062.0, "reward": 1.1013752222061157, "reward_std": 0.32002100348472595, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6246234178543091, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9548437595367432, "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": 701.0, "completions/max_terminated_length": 701.0, "completions/mean_length": 356.9375, "completions/mean_terminated_length": 356.9375, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.0208, "grad_norm": 0.06536313891410828, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15298682.0, "reward": 0.9743213057518005, "reward_std": 0.4704718291759491, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4955156445503235, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 714.0, "completions/mean_length": 359.390625, "completions/mean_terminated_length": 348.8412780761719, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 1.0216, "grad_norm": 0.05333084613084793, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15338355.0, "reward": 0.7093064188957214, "reward_std": 0.2049013376235962, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30923593044281006, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7729687690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 418.59375, "completions/mean_terminated_length": 408.9841613769531, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 1.0224, "grad_norm": 0.04256478324532509, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15380393.0, "reward": 0.9115470051765442, "reward_std": 0.3410246968269348, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.49496719241142273, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8721874952316284, "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": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 381.609375, "completions/mean_terminated_length": 381.609375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 1.0232, "grad_norm": 0.05282418429851532, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15422960.0, "reward": 1.0202500820159912, "reward_std": 0.26850804686546326, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.6342484354972839, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8142187595367432, "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": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 375.296875, "completions/mean_terminated_length": 375.296875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.024, "grad_norm": 0.046594515442848206, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15464691.0, "reward": 0.9427322149276733, "reward_std": 0.3763318359851837, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47921252250671387, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8192187547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 735.0, "completions/mean_length": 427.546875, "completions/mean_terminated_length": 408.3064270019531, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.0248, "grad_norm": 0.04241950809955597, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15508118.0, "reward": 0.8450456261634827, "reward_std": 0.2906612157821655, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.42446404695510864, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8645312786102295, "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": 587.0, "completions/max_terminated_length": 587.0, "completions/mean_length": 350.625, "completions/mean_terminated_length": 350.625, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 1.0256, "grad_norm": 0.045778654515743256, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15546046.0, "reward": 1.0432275533676147, "reward_std": 0.24005407094955444, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6020781397819519, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9175000190734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 591.0, "completions/mean_length": 368.125, "completions/mean_terminated_length": 357.71429443359375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.0264, "grad_norm": 0.058314330875873566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15585750.0, "reward": 0.9423236846923828, "reward_std": 0.3953005373477936, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4783953130245209, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9410937428474426, "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": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 414.4375, "completions/mean_terminated_length": 414.4375, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.0272, "grad_norm": 0.06670884788036346, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15627714.0, "reward": 1.0512197017669678, "reward_std": 0.45513108372688293, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5711874961853027, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9540624618530273, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 500.015625, "completions/mean_terminated_length": 491.6984558105469, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 1.028, "grad_norm": 0.0324665792286396, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15675955.0, "reward": 0.9579673409461975, "reward_std": 0.24387867748737335, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5721827745437622, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8123437166213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 466.421875, "completions/mean_terminated_length": 457.5714416503906, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 1.0288, "grad_norm": 0.03803452476859093, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15722190.0, "reward": 1.0462743043899536, "reward_std": 0.3675340414047241, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6394218802452087, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.839062511920929, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 450.046875, "completions/mean_terminated_length": 379.5614013671875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.0296, "grad_norm": 0.06279318779706955, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15767025.0, "reward": 0.6818150281906128, "reward_std": 0.35974860191345215, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.2855031192302704, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8276562094688416, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 495.015625, "completions/mean_terminated_length": 486.61907958984375, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 1.0304, "grad_norm": 0.04572480544447899, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15814178.0, "reward": 1.1306877136230469, "reward_std": 0.4206940531730652, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6519984602928162, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9042187929153442, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 573.578125, "completions/mean_terminated_length": 526.9827270507812, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.0312, "grad_norm": 0.05971655249595642, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15866615.0, "reward": 0.5744227766990662, "reward_std": 0.3567269444465637, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.21134373545646667, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.8189062476158142, "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": 593.0, "completions/max_terminated_length": 593.0, "completions/mean_length": 242.59375, "completions/mean_terminated_length": 242.59375, "completions/min_length": 86.0, "completions/min_terminated_length": 86.0, "epoch": 1.032, "grad_norm": 0.06491947919130325, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15896157.0, "reward": 0.8492714762687683, "reward_std": 0.12659798562526703, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.3704156279563904, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9700000286102295, "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": 618.0, "completions/max_terminated_length": 618.0, "completions/mean_length": 339.546875, "completions/mean_terminated_length": 339.546875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.0328, "grad_norm": 0.0629461333155632, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 15933632.0, "reward": 1.1314690113067627, "reward_std": 0.40147799253463745, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.684810996055603, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8957812786102295, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 504.796875, "completions/mean_terminated_length": 479.26226806640625, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 1.0336, "grad_norm": 0.06385250389575958, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 15982147.0, "reward": 0.9011431932449341, "reward_std": 0.3840924799442291, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4897843599319458, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8506250381469727, "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": 0.0, "completions/max_length": 952.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 388.46875, "completions/mean_terminated_length": 388.46875, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "epoch": 1.0344, "grad_norm": 0.028928296640515327, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16022705.0, "reward": 0.710858941078186, "reward_std": 0.1905198097229004, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.2498406320810318, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500810623169, "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": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 366.625, "completions/mean_terminated_length": 366.625, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 1.0352, "grad_norm": 0.06662557274103165, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16062329.0, "reward": 0.9564533829689026, "reward_std": 0.2983527183532715, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4754047095775604, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.944531261920929, "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": 845.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 432.828125, "completions/mean_terminated_length": 432.828125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.036, "grad_norm": 0.03651294857263565, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16107806.0, "reward": 1.1633479595184326, "reward_std": 0.32949599623680115, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6860687732696533, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9550000429153442, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 415.09375, "completions/mean_terminated_length": 405.4285888671875, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.0368, "grad_norm": 0.04794280603528023, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16150452.0, "reward": 0.8498360514640808, "reward_std": 0.3096178472042084, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.41842031478881836, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8989062309265137, "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": 832.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 358.421875, "completions/mean_terminated_length": 358.421875, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 1.0376, "grad_norm": 0.05998740345239639, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16189295.0, "reward": 1.0308377742767334, "reward_std": 0.38070791959762573, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5616734623908997, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9448437690734863, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 451.625, "completions/mean_terminated_length": 423.4753723144531, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 1.0384, "grad_norm": 0.05892918258905411, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16234375.0, "reward": 0.7722165584564209, "reward_std": 0.3349781632423401, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.38818126916885376, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.807812511920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 447.765625, "completions/mean_terminated_length": 438.61907958984375, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.0392, "grad_norm": 0.04173288866877556, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16279944.0, "reward": 0.9886924028396606, "reward_std": 0.3194853365421295, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5398828387260437, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9167187213897705, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 484.4375, "completions/mean_terminated_length": 457.901611328125, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 1.04, "grad_norm": 0.04772281274199486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16328292.0, "reward": 0.6617742776870728, "reward_std": 0.3002663850784302, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.3391718566417694, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7659374475479126, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 526.53125, "completions/mean_terminated_length": 502.0655517578125, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.0408, "grad_norm": 0.04478205740451813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16378150.0, "reward": 0.803078293800354, "reward_std": 0.4219137728214264, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.3717796802520752, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8879687786102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 463.96875, "completions/mean_terminated_length": 455.0793762207031, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 1.0416, "grad_norm": 0.025595257058739662, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16422932.0, "reward": 0.962425172328949, "reward_std": 0.2572762966156006, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4873484671115875, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.947029709815979, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 676.0, "completions/mean_length": 440.453125, "completions/mean_terminated_length": 431.19049072265625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 1.0424, "grad_norm": 0.05393434315919876, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16466241.0, "reward": 0.9865025281906128, "reward_std": 0.46315258741378784, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5355031490325928, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9112499952316284, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 909.0, "completions/mean_length": 511.234375, "completions/mean_terminated_length": 494.69354248046875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.0432, "grad_norm": 0.030918356031179428, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16515328.0, "reward": 0.8172595500946045, "reward_std": 0.302118182182312, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.43139219284057617, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.827343761920929, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 481.453125, "completions/mean_terminated_length": 454.7704772949219, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.044, "grad_norm": 0.045021768659353256, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16563389.0, "reward": 1.0511713027954102, "reward_std": 0.3766826093196869, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6179656386375427, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.875, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 426.5, "completions/mean_terminated_length": 397.11474609375, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.0448, "grad_norm": 0.09257517009973526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16606749.0, "reward": 0.8401229381561279, "reward_std": 0.3800368309020996, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41461873054504395, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8768749833106995, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 496.4375, "completions/mean_terminated_length": 488.0635070800781, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.0456, "grad_norm": 0.04106690362095833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16655417.0, "reward": 0.9942877292633057, "reward_std": 0.4192064106464386, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.535448431968689, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9239062070846558, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 343.625, "completions/mean_terminated_length": 310.1639099121094, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 1.0464, "grad_norm": 0.06859569251537323, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16692289.0, "reward": 0.8141947984695435, "reward_std": 0.3468829393386841, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4096375107765198, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8496875166893005, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 454.15625, "completions/mean_terminated_length": 445.11114501953125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.0472, "grad_norm": 0.04522018879652023, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16738299.0, "reward": 0.9368300437927246, "reward_std": 0.3323354721069336, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5142828226089478, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8817187547683716, "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": 963.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 447.0625, "completions/mean_terminated_length": 447.0625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.048, "grad_norm": 0.059221673756837845, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16782495.0, "reward": 1.146282434463501, "reward_std": 0.3020627796649933, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6831874847412109, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9318749904632568, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 322.453125, "completions/mean_terminated_length": 299.82257080078125, "completions/min_length": 105.0, "completions/min_terminated_length": 105.0, "epoch": 1.0488, "grad_norm": 0.07984521985054016, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16819036.0, "reward": 0.8460075259208679, "reward_std": 0.3984881639480591, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.3951374888420105, "rewards/confidence_one_or_zero": 0.34375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9140625, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 514.5, "completions/mean_terminated_length": 506.4127197265625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 1.0496, "grad_norm": 0.0429844968020916, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16867756.0, "reward": 0.8561657667160034, "reward_std": 0.3905707597732544, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4467046856880188, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.844531238079071, "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": 0.0, "completions/max_length": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 449.265625, "completions/mean_terminated_length": 449.265625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.0504, "grad_norm": 0.04675872623920441, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16913693.0, "reward": 0.827914297580719, "reward_std": 0.3412652015686035, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.37457653880119324, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.928906261920929, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 319.515625, "completions/mean_terminated_length": 296.7903137207031, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "epoch": 1.0512, "grad_norm": 0.05379561707377434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 16950254.0, "reward": 1.1061253547668457, "reward_std": 0.38861989974975586, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6809984445571899, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7517187595367432, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 532.078125, "completions/mean_terminated_length": 516.2096557617188, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 1.052, "grad_norm": 0.04847264289855957, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 16999987.0, "reward": 0.915885329246521, "reward_std": 0.3533000946044922, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4880187511444092, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8885937929153442, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 559.734375, "completions/mean_terminated_length": 511.7069091796875, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.0528, "grad_norm": 0.05923382565379143, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17051762.0, "reward": 1.0656015872955322, "reward_std": 0.4646647870540619, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6624516248703003, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8204687833786011, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 507.28125, "completions/mean_terminated_length": 453.82757568359375, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 1.0536, "grad_norm": 0.03419993817806244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17099556.0, "reward": 0.9774494171142578, "reward_std": 0.4249767065048218, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5486469268798828, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8496874570846558, "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": 0.265625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 643.734375, "completions/mean_terminated_length": 506.19146728515625, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.0544, "grad_norm": 0.04615318775177002, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17158883.0, "reward": 0.7000725269317627, "reward_std": 0.41993871331214905, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.40014374256134033, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.734375, "rewards/mean_confidence_reward": 0.6225000023841858, "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": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 390.453125, "completions/mean_terminated_length": 390.453125, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 1.0552, "grad_norm": 0.04427509382367134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17199488.0, "reward": 1.161635398864746, "reward_std": 0.26891008019447327, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6826437711715698, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9573421478271484, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 593.625, "completions/mean_terminated_length": 557.1525268554688, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 1.056, "grad_norm": 0.05014043673872948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17253320.0, "reward": 1.115120530128479, "reward_std": 0.4600270390510559, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6989890336990356, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8470312356948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 413.40625, "completions/mean_terminated_length": 403.71429443359375, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 1.0568, "grad_norm": 0.07288504391908646, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17295170.0, "reward": 1.174200177192688, "reward_std": 0.4579569697380066, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.7233984470367432, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8929687738418579, "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": 933.0, "completions/mean_length": 565.9375, "completions/mean_terminated_length": 535.4000244140625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.0576, "grad_norm": 0.03895450755953789, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17346478.0, "reward": 0.6347421407699585, "reward_std": 0.28108566999435425, "rewards/accuracy_reward": 0.03125, "rewards/brier_reward": 0.3319827914237976, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7017187476158142, "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": 0.0, "completions/max_length": 929.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 410.515625, "completions/mean_terminated_length": 410.515625, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 1.0584, "grad_norm": 0.03764556348323822, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17389151.0, "reward": 1.142356514930725, "reward_std": 0.2864346504211426, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.7065858840942383, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8920312523841858, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 615.390625, "completions/mean_terminated_length": 573.1206665039062, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 1.0592, "grad_norm": 0.043054286390542984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17444280.0, "reward": 0.7763804197311401, "reward_std": 0.42460691928863525, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.4590093791484833, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.645312488079071, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 382.65625, "completions/mean_terminated_length": 372.4762268066406, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.06, "grad_norm": 0.07135516405105591, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17484978.0, "reward": 1.170868158340454, "reward_std": 0.4574054479598999, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7167344093322754, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.91734379529953, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 797.0, "completions/mean_length": 409.640625, "completions/mean_terminated_length": 399.888916015625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.0608, "grad_norm": 0.04979077726602554, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17526939.0, "reward": 0.8612274527549744, "reward_std": 0.31940335035324097, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.48807811737060547, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7890625, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 452.859375, "completions/mean_terminated_length": 443.7936706542969, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.0616, "grad_norm": 0.059418998658657074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17572002.0, "reward": 1.0543065071105957, "reward_std": 0.3984832465648651, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.5929859280586243, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9401562809944153, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 400.03125, "completions/mean_terminated_length": 390.12701416015625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.0624, "grad_norm": 0.0378069244325161, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17614900.0, "reward": 0.7796987295150757, "reward_std": 0.2814478278160095, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.30939531326293945, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9401562213897705, "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": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 344.03125, "completions/mean_terminated_length": 344.03125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 1.0632, "grad_norm": 0.043250348418951035, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17654214.0, "reward": 1.2754220962524414, "reward_std": 0.29600462317466736, "rewards/accuracy_reward": 0.765625, "rewards/brier_reward": 0.7852171659469604, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.960156261920929, "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": 0.0, "completions/max_length": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 383.640625, "completions/mean_terminated_length": 383.640625, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.064, "grad_norm": 0.05229544639587402, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17694863.0, "reward": 0.9675040245056152, "reward_std": 0.322858989238739, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5287562608718872, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8659374713897705, "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": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 557.796875, "completions/mean_terminated_length": 557.796875, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.0648, "grad_norm": 0.018496008589863777, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17747954.0, "reward": 0.696831464767456, "reward_std": 0.19762197136878967, "rewards/accuracy_reward": 0.03125, "rewards/brier_reward": 0.36241093277931213, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.74609375, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 611.734375, "completions/mean_terminated_length": 552.8392944335938, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.0656, "grad_norm": 0.04687870293855667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17805393.0, "reward": 0.6575954556465149, "reward_std": 0.4338068664073944, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.28393906354904175, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.8051562309265137, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 601.453125, "completions/mean_terminated_length": 557.7413940429688, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 1.0664, "grad_norm": 0.0357605516910553, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17860574.0, "reward": 0.6465361714363098, "reward_std": 0.3742744028568268, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.27744531631469727, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7779687643051147, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 500.546875, "completions/mean_terminated_length": 446.39654541015625, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.0672, "grad_norm": 0.05852751061320305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17910449.0, "reward": 0.8521103262901306, "reward_std": 0.4373372793197632, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4542187452316284, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.826718807220459, "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": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 386.703125, "completions/mean_terminated_length": 386.703125, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.068, "grad_norm": 0.05936021730303764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 17951646.0, "reward": 0.9690767526626587, "reward_std": 0.37014293670654297, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5162765979766846, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9142187833786011, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 630.0, "completions/mean_length": 441.875, "completions/mean_terminated_length": 413.2458801269531, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 1.0688, "grad_norm": 0.05399996414780617, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 17995542.0, "reward": 0.9757579565048218, "reward_std": 0.4929189682006836, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5296390056610107, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8938906192779541, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 463.59375, "completions/mean_terminated_length": 454.69842529296875, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.0695999999999999, "grad_norm": 0.04496889188885689, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18041500.0, "reward": 1.061428427696228, "reward_std": 0.4022926986217499, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6228547096252441, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8823437690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 488.96875, "completions/mean_terminated_length": 480.4762268066406, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.0704, "grad_norm": 0.04631191864609718, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18089914.0, "reward": 0.7581509351730347, "reward_std": 0.3539796471595764, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.3287999927997589, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9009374976158142, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 496.5, "completions/mean_terminated_length": 488.12701416015625, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 1.0712, "grad_norm": 0.04752276465296745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18137386.0, "reward": 0.742505669593811, "reward_std": 0.24549299478530884, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.2975093722343445, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9099999666213989, "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": 0.21875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 597.09375, "completions/mean_terminated_length": 477.55999755859375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.072, "grad_norm": 0.024955635890364647, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18191824.0, "reward": 0.45913517475128174, "reward_std": 0.22071129083633423, "rewards/accuracy_reward": 0.015625, "rewards/brier_reward": 0.13701874017715454, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.765625, "rewards/mean_confidence_reward": 0.6806250214576721, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 468.03125, "completions/mean_terminated_length": 459.2063903808594, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 1.0728, "grad_norm": 0.08959347009658813, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18240546.0, "reward": 0.9134392738342285, "reward_std": 0.4385979473590851, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.45187655091285706, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8973437547683716, "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": 1020.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 542.6875, "completions/mean_terminated_length": 542.6875, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 1.0735999999999999, "grad_norm": 0.04941180720925331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18292094.0, "reward": 0.75520259141922, "reward_std": 0.3510030508041382, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3385280966758728, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9217187166213989, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 614.484375, "completions/mean_terminated_length": 572.1206665039062, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 1.0744, "grad_norm": 0.029695790261030197, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18347085.0, "reward": 0.7493687868118286, "reward_std": 0.31493353843688965, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4362359642982483, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6906249523162842, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 506.34375, "completions/mean_terminated_length": 498.12701416015625, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.0752, "grad_norm": 0.035048436373472214, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18397379.0, "reward": 0.7565760612487793, "reward_std": 0.2956218719482422, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.31002500653266907, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9121875166893005, "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": 968.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 398.984375, "completions/mean_terminated_length": 398.984375, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.076, "grad_norm": 0.03628474101424217, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18438738.0, "reward": 0.8700441122055054, "reward_std": 0.21614152193069458, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4432109594345093, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8689062595367432, "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": 916.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 388.203125, "completions/mean_terminated_length": 388.203125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.0768, "grad_norm": 0.06211915612220764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18480607.0, "reward": 0.9378463625907898, "reward_std": 0.4172651767730713, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.46944063901901245, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9350000023841858, "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": 647.0, "completions/max_terminated_length": 647.0, "completions/mean_length": 304.84375, "completions/mean_terminated_length": 304.84375, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 1.0776, "grad_norm": 0.08639606088399887, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18516133.0, "reward": 0.9170103073120117, "reward_std": 0.3584516942501068, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.459018737077713, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9318749904632568, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 830.0, "completions/mean_length": 424.484375, "completions/mean_terminated_length": 394.9999694824219, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 1.0784, "grad_norm": 0.06477797776460648, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18559012.0, "reward": 0.6727173924446106, "reward_std": 0.32001906633377075, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.2516828179359436, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8543750047683716, "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": 0.0, "completions/max_length": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 469.671875, "completions/mean_terminated_length": 469.671875, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.0792, "grad_norm": 0.04977261275053024, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18604975.0, "reward": 0.8772860765457153, "reward_std": 0.319516122341156, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.5045703053474426, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7984374761581421, "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": 870.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 475.9375, "completions/mean_terminated_length": 475.9375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.08, "grad_norm": 0.044978369027376175, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18651803.0, "reward": 0.8781086206436157, "reward_std": 0.30706265568733215, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.5062156319618225, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8075000047683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 411.40625, "completions/mean_terminated_length": 401.68255615234375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 1.0808, "grad_norm": 0.06599236279726028, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18692997.0, "reward": 0.9645105600357056, "reward_std": 0.480766236782074, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5071437358856201, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9281250238418579, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 485.15625, "completions/mean_terminated_length": 476.60321044921875, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 1.0816, "grad_norm": 0.015410560183227062, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18741071.0, "reward": 1.1519749164581299, "reward_std": 0.14340518414974213, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.8039484620094299, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.6792187690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 501.0625, "completions/mean_terminated_length": 484.19354248046875, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 1.0824, "grad_norm": 0.030761979520320892, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18789971.0, "reward": 0.9942179918289185, "reward_std": 0.34439200162887573, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.6446843147277832, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7450000047683716, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 538.796875, "completions/mean_terminated_length": 448.9444580078125, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.0832, "grad_norm": 0.053454089909791946, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18840310.0, "reward": 0.8736906051635742, "reward_std": 0.4743279218673706, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.513004720211029, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.7332812547683716, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 646.078125, "completions/mean_terminated_length": 606.9827270507812, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 1.084, "grad_norm": 0.023764628916978836, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18897963.0, "reward": 0.8542101383209229, "reward_std": 0.25589263439178467, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5052937269210815, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7435937523841858, "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": 859.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 462.5625, "completions/mean_terminated_length": 462.5625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 1.0848, "grad_norm": 0.04687805846333504, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 18943839.0, "reward": 0.9747072458267212, "reward_std": 0.3231302499771118, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5275375247001648, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9159374833106995, "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": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 394.296875, "completions/mean_terminated_length": 394.296875, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 1.0856, "grad_norm": 0.06735389679670334, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 18985986.0, "reward": 0.8376986384391785, "reward_std": 0.3627094030380249, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.39414530992507935, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9267187714576721, "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": 0.0, "completions/max_length": 913.0, "completions/max_terminated_length": 913.0, "completions/mean_length": 479.609375, "completions/mean_terminated_length": 479.609375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 1.0864, "grad_norm": 0.03991466015577316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19034969.0, "reward": 1.1297290325164795, "reward_std": 0.2271406650543213, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6657062768936157, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9420312643051147, "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": 1011.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 534.046875, "completions/mean_terminated_length": 534.046875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.0872, "grad_norm": 0.04145023971796036, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19086732.0, "reward": 1.1335102319717407, "reward_std": 0.36054348945617676, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.7357687950134277, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8009375333786011, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 505.796875, "completions/mean_terminated_length": 489.08062744140625, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 1.088, "grad_norm": 0.028860021382570267, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19136607.0, "reward": 0.9299150705337524, "reward_std": 0.2287282943725586, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.5785781145095825, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7589062452316284, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 506.375, "completions/mean_terminated_length": 498.15875244140625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.0888, "grad_norm": 0.03664294257760048, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19185495.0, "reward": 1.104541540145874, "reward_std": 0.35403743386268616, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.72470623254776, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7993749976158142, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 529.421875, "completions/mean_terminated_length": 505.09832763671875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.0896, "grad_norm": 0.04668673500418663, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19234370.0, "reward": 0.7600728273391724, "reward_std": 0.4106138348579407, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.4263937473297119, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7485937476158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 539.890625, "completions/mean_terminated_length": 524.274169921875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 1.0904, "grad_norm": 0.0504380464553833, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19287659.0, "reward": 0.7847118377685547, "reward_std": 0.43727320432662964, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.3819218575954437, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8762499690055847, "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": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 425.328125, "completions/mean_terminated_length": 425.328125, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.0912, "grad_norm": 0.045457132160663605, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19330736.0, "reward": 0.9987165927886963, "reward_std": 0.3549611270427704, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5443062782287598, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9193750619888306, "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": 912.0, "completions/max_terminated_length": 912.0, "completions/mean_length": 460.875, "completions/mean_terminated_length": 460.875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.092, "grad_norm": 0.05584071949124336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19377000.0, "reward": 0.8780993819236755, "reward_std": 0.37206944823265076, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.45932188630104065, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8765624761581421, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 450.328125, "completions/mean_terminated_length": 422.1147155761719, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.0928, "grad_norm": 0.058213960379362106, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19422173.0, "reward": 1.1543954610824585, "reward_std": 0.47209158539772034, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.715039074420929, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.87890625, "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": 574.0, "completions/max_terminated_length": 574.0, "completions/mean_length": 330.765625, "completions/mean_terminated_length": 330.765625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.0936, "grad_norm": 0.04537295177578926, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19459790.0, "reward": 1.0372118949890137, "reward_std": 0.29287153482437134, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5744218826293945, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9378125667572021, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 438.109375, "completions/mean_terminated_length": 428.8095397949219, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.0944, "grad_norm": 0.06731703877449036, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19504741.0, "reward": 0.977347731590271, "reward_std": 0.37834471464157104, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5953187942504883, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8031250238418579, "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": 0.0, "completions/max_length": 767.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 360.328125, "completions/mean_terminated_length": 360.328125, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 1.0952, "grad_norm": 0.07957465201616287, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19543738.0, "reward": 1.1021925210952759, "reward_std": 0.42629408836364746, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6262578368186951, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9535937309265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 562.953125, "completions/mean_terminated_length": 548.0806274414062, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 1.096, "grad_norm": 0.036721568554639816, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19597703.0, "reward": 0.6429744958877563, "reward_std": 0.312640905380249, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.20782187581062317, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9021874666213989, "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": 811.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 409.734375, "completions/mean_terminated_length": 409.734375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.0968, "grad_norm": 0.042327579110860825, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19640230.0, "reward": 1.085676670074463, "reward_std": 0.37537604570388794, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6244765520095825, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9417187571525574, "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": 911.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 462.390625, "completions/mean_terminated_length": 462.390625, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 1.0976, "grad_norm": 0.07766316086053848, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19687247.0, "reward": 1.1408385038375854, "reward_std": 0.41323819756507874, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.7035499811172485, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9164062738418579, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 549.015625, "completions/mean_terminated_length": 490.6842041015625, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 1.0984, "grad_norm": 0.047968942672014236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19738368.0, "reward": 0.8277899026870728, "reward_std": 0.42429378628730774, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4837031364440918, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6765625476837158, "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": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 434.359375, "completions/mean_terminated_length": 434.359375, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.0992, "grad_norm": 0.03650001063942909, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19782887.0, "reward": 1.0521199703216553, "reward_std": 0.2373497188091278, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5729874968528748, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9634374976158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 522.828125, "completions/mean_terminated_length": 506.6612854003906, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.1, "grad_norm": 0.04744623228907585, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19832508.0, "reward": 0.8725666403770447, "reward_std": 0.4485422968864441, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.417006254196167, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9206249713897705, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 470.59375, "completions/mean_terminated_length": 443.37701416015625, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.1008, "grad_norm": 0.047631893306970596, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19878770.0, "reward": 0.7414892315864563, "reward_std": 0.35297954082489014, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3579765558242798, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8160937428474426, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 524.46875, "completions/mean_terminated_length": 508.3548278808594, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.1016, "grad_norm": 0.03891067951917648, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 19929424.0, "reward": 0.8009282350540161, "reward_std": 0.37827762961387634, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.49247971177101135, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.6740624904632568, "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": 836.0, "completions/max_terminated_length": 836.0, "completions/mean_length": 407.90625, "completions/mean_terminated_length": 407.90625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 1.1024, "grad_norm": 0.037979856133461, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 19971210.0, "reward": 1.3273236751556396, "reward_std": 0.30233198404312134, "rewards/accuracy_reward": 0.8125, "rewards/brier_reward": 0.8421453237533569, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.946093738079071, "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": 0.0, "completions/max_length": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 466.546875, "completions/mean_terminated_length": 466.546875, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.1032, "grad_norm": 0.07446563988924026, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20017837.0, "reward": 1.1550650596618652, "reward_std": 0.46351271867752075, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6851280927658081, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9493734240531921, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 559.15625, "completions/mean_terminated_length": 551.77783203125, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 1.104, "grad_norm": 0.03952465206384659, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20068727.0, "reward": 0.8603932857513428, "reward_std": 0.39225104451179504, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.39265936613082886, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9440624713897705, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 572.15625, "completions/mean_terminated_length": 533.8643798828125, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.1048, "grad_norm": 0.04581112414598465, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20122561.0, "reward": 0.6898571848869324, "reward_std": 0.37505975365638733, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.3015874922275543, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8125, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 543.5, "completions/mean_terminated_length": 511.4667053222656, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 1.1056, "grad_norm": 0.032558541744947433, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20173777.0, "reward": 0.627927303314209, "reward_std": 0.25172045826911926, "rewards/accuracy_reward": 0.0, "rewards/brier_reward": 0.3652281165122986, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6317187547683716, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 534.53125, "completions/mean_terminated_length": 474.4210510253906, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.1064, "grad_norm": 0.04932728409767151, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20224531.0, "reward": 0.9919494390487671, "reward_std": 0.4123438000679016, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5776468515396118, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8278124928474426, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 808.0, "completions/mean_length": 596.390625, "completions/mean_terminated_length": 535.3035888671875, "completions/min_length": 350.0, "completions/min_terminated_length": 350.0, "epoch": 1.1072, "grad_norm": 0.03758999705314636, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20278972.0, "reward": 1.0347172021865845, "reward_std": 0.439627468585968, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6475578546524048, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7282812595367432, "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": 927.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 463.296875, "completions/mean_terminated_length": 463.296875, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 1.108, "grad_norm": 0.050214435905218124, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20324623.0, "reward": 0.9961955547332764, "reward_std": 0.2897912561893463, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5392640829086304, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9214062690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 550.71875, "completions/mean_terminated_length": 535.4515991210938, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.1088, "grad_norm": 0.037339046597480774, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20375261.0, "reward": 0.8707946538925171, "reward_std": 0.32828447222709656, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4915875196456909, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.793749988079071, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 608.953125, "completions/mean_terminated_length": 588.5409545898438, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 1.1096, "grad_norm": 0.044868819415569305, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20430666.0, "reward": 0.852321982383728, "reward_std": 0.3182886838912964, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4233921766281128, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8785937428474426, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 628.0, "completions/mean_length": 423.921875, "completions/mean_terminated_length": 394.4098205566406, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.1104, "grad_norm": 0.05722685158252716, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20474501.0, "reward": 0.9453814625740051, "reward_std": 0.4591296911239624, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.48451095819473267, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9142187833786011, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 521.171875, "completions/mean_terminated_length": 504.95159912109375, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.1112, "grad_norm": 0.0455741286277771, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20524800.0, "reward": 0.8707611560821533, "reward_std": 0.335939884185791, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.47589531540870667, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8082812428474426, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 805.0, "completions/mean_length": 432.96875, "completions/mean_terminated_length": 423.5873107910156, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.112, "grad_norm": 0.050487712025642395, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20570798.0, "reward": 0.717633068561554, "reward_std": 0.3295551836490631, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.34151405096054077, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7939062714576721, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 752.0, "completions/mean_length": 471.515625, "completions/mean_terminated_length": 444.34423828125, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.1128, "grad_norm": 0.059437867254018784, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20616751.0, "reward": 0.920537531375885, "reward_std": 0.4222177565097809, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.6066984534263611, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6926562786102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 397.875, "completions/mean_terminated_length": 387.9365234375, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 1.1136, "grad_norm": 0.055011458694934845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20658071.0, "reward": 1.1459846496582031, "reward_std": 0.40816402435302734, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6982171535491943, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.897656261920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 517.640625, "completions/mean_terminated_length": 509.60321044921875, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.1144, "grad_norm": 0.046840257942676544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20707056.0, "reward": 0.9846691489219666, "reward_std": 0.36392056941986084, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5162109136581421, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9401562809944153, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 485.71875, "completions/mean_terminated_length": 459.2458801269531, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 1.1152, "grad_norm": 0.05534703657031059, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20753774.0, "reward": 0.7738009691238403, "reward_std": 0.37374773621559143, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.29760000109672546, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8215625286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 681.0, "completions/mean_length": 319.84375, "completions/mean_terminated_length": 308.66668701171875, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.116, "grad_norm": 0.09329301863908768, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20789780.0, "reward": 0.8268276453018188, "reward_std": 0.42332547903060913, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.3880281448364258, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9095312356948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 510.484375, "completions/mean_terminated_length": 502.3333740234375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 1.1168, "grad_norm": 0.03838421776890755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20838243.0, "reward": 0.7672034502029419, "reward_std": 0.4009307026863098, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3156546950340271, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.912656307220459, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 387.671875, "completions/mean_terminated_length": 377.5714416503906, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 1.1176, "grad_norm": 0.05309082567691803, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20879950.0, "reward": 1.0688657760620117, "reward_std": 0.39549776911735535, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.5752297043800354, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8695312738418579, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 531.4375, "completions/mean_terminated_length": 489.6949157714844, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.1184, "grad_norm": 0.036942094564437866, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 20931722.0, "reward": 0.9079657196998596, "reward_std": 0.367425799369812, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.503429651260376, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8073437213897705, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 435.984375, "completions/mean_terminated_length": 407.0655517578125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 1.1192, "grad_norm": 0.0520176887512207, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 20975273.0, "reward": 1.1575665473937988, "reward_std": 0.2599279582500458, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7057563066482544, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8890625238418579, "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": 729.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 332.78125, "completions/mean_terminated_length": 332.78125, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.12, "grad_norm": 0.06367902457714081, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21012859.0, "reward": 0.9021362066268921, "reward_std": 0.3688526451587677, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.42927032709121704, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9640625715255737, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 417.0, "completions/mean_terminated_length": 407.3651123046875, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 1.1208, "grad_norm": 0.05482793226838112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21056203.0, "reward": 1.152010440826416, "reward_std": 0.4642331004142761, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6790187358856201, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9507812261581421, "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": 0.0, "completions/max_length": 969.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 480.59375, "completions/mean_terminated_length": 480.59375, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 1.1216, "grad_norm": 0.04670625180006027, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21102353.0, "reward": 0.8623619675636292, "reward_std": 0.30237284302711487, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.41222187876701355, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9146875143051147, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 403.984375, "completions/mean_terminated_length": 394.14288330078125, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 1.1224, "grad_norm": 0.06670387834310532, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21144016.0, "reward": 0.9768252372741699, "reward_std": 0.4321291148662567, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5005234479904175, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9535781145095825, "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": 0.0, "completions/max_length": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 494.9375, "completions/mean_terminated_length": 494.9375, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.1232, "grad_norm": 0.06568682938814163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21192972.0, "reward": 0.9941314458847046, "reward_std": 0.39237338304519653, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5663859844207764, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8770312070846558, "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": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 397.765625, "completions/mean_terminated_length": 397.765625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.124, "grad_norm": 0.07629229128360748, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21234157.0, "reward": 1.0574760437011719, "reward_std": 0.47040310502052307, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5837000012397766, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9635937213897705, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 840.0, "completions/mean_length": 475.546875, "completions/mean_terminated_length": 457.8548278808594, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.1248, "grad_norm": 0.06059170141816139, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21281232.0, "reward": 0.8247172832489014, "reward_std": 0.40973541140556335, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.41505780816078186, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8564062714576721, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 532.65625, "completions/mean_terminated_length": 524.857177734375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 1.1256, "grad_norm": 0.04372258856892586, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21333866.0, "reward": 0.8219846487045288, "reward_std": 0.3806360960006714, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4564671814441681, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8274999856948853, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 520.953125, "completions/mean_terminated_length": 496.21307373046875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.1264, "grad_norm": 0.04538174718618393, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21383943.0, "reward": 0.7239589691162109, "reward_std": 0.362635999917984, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.29166561365127563, "rewards/confidence_one_or_zero": 0.28125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8989062309265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 534.34375, "completions/mean_terminated_length": 518.5484008789062, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 1.1272, "grad_norm": 0.046059589833021164, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21434269.0, "reward": 1.0152735710144043, "reward_std": 0.43569692969322205, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5774203538894653, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9118750095367432, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 490.015625, "completions/mean_terminated_length": 463.7540588378906, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.1280000000000001, "grad_norm": 0.0522906668484211, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21481710.0, "reward": 0.8815876245498657, "reward_std": 0.4742578864097595, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4662984311580658, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8482655882835388, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 504.828125, "completions/mean_terminated_length": 470.2167053222656, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.1288, "grad_norm": 0.04687722027301788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21529539.0, "reward": 1.033628225326538, "reward_std": 0.4129236936569214, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6453796625137329, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8023437261581421, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 397.796875, "completions/mean_terminated_length": 377.5967712402344, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.1296, "grad_norm": 0.059213075786828995, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21571366.0, "reward": 0.9077867865562439, "reward_std": 0.378963440656662, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4874468743801117, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.86781245470047, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 886.0, "completions/mean_length": 485.0, "completions/mean_terminated_length": 476.4444580078125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.1304, "grad_norm": 0.04591954126954079, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21619654.0, "reward": 0.8932963013648987, "reward_std": 0.3770716190338135, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4740906357765198, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8718750476837158, "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": 841.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 463.390625, "completions/mean_terminated_length": 463.390625, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 1.1312, "grad_norm": 0.03897688537836075, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21665407.0, "reward": 0.7953914403915405, "reward_std": 0.26554083824157715, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.5282812714576721, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.6000000238418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 503.640625, "completions/mean_terminated_length": 486.8548278808594, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.1320000000000001, "grad_norm": 0.05282946303486824, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21714040.0, "reward": 0.8937087059020996, "reward_std": 0.4120439887046814, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.49054062366485596, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8392187356948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 509.203125, "completions/mean_terminated_length": 501.0317687988281, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 1.1328, "grad_norm": 0.05171431973576546, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21764197.0, "reward": 0.8122378587722778, "reward_std": 0.4289678633213043, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.3744734525680542, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9076562523841858, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 453.71875, "completions/mean_terminated_length": 444.66668701171875, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.1336, "grad_norm": 0.04712458327412605, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21810579.0, "reward": 0.8971558809280396, "reward_std": 0.40438133478164673, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49743497371673584, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8553390502929688, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 510.0, "completions/mean_terminated_length": 484.7212829589844, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 1.1344, "grad_norm": 0.038255058228969574, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21860483.0, "reward": 0.7185595035552979, "reward_std": 0.3129946291446686, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.32774218916893005, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8048437237739563, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 455.53125, "completions/mean_terminated_length": 446.5079650878906, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.1352, "grad_norm": 0.06314196437597275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 21906053.0, "reward": 0.9543431997299194, "reward_std": 0.41141560673713684, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4868093729019165, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.942812442779541, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 655.109375, "completions/mean_terminated_length": 623.8474731445312, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 1.1360000000000001, "grad_norm": 0.03261676803231239, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 21963196.0, "reward": 0.627572774887085, "reward_std": 0.3131600618362427, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.25514376163482666, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.818906307220459, "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": 947.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 510.03125, "completions/mean_terminated_length": 510.03125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 1.1368, "grad_norm": 0.04515283182263374, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22013486.0, "reward": 0.8654744625091553, "reward_std": 0.35447901487350464, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4340718686580658, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8967187404632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 911.0, "completions/mean_length": 498.015625, "completions/mean_terminated_length": 489.66668701171875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.1376, "grad_norm": 0.028210777789354324, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22062975.0, "reward": 0.7753656506538391, "reward_std": 0.24265874922275543, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.4569796919822693, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.681093692779541, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 615.78125, "completions/mean_terminated_length": 595.7048950195312, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 1.1384, "grad_norm": 0.043187256902456284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22121793.0, "reward": 0.7552065849304199, "reward_std": 0.3697644770145416, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.35416093468666077, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8456250429153442, "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": 994.0, "completions/mean_length": 568.484375, "completions/mean_terminated_length": 538.11669921875, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.1392, "grad_norm": 0.05028944090008736, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22174032.0, "reward": 0.8357251882553101, "reward_std": 0.4350491762161255, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4058234691619873, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8757812976837158, "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": 569.0, "completions/max_terminated_length": 569.0, "completions/mean_length": 385.15625, "completions/mean_terminated_length": 385.15625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.1400000000000001, "grad_norm": 0.040851712226867676, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22215066.0, "reward": 1.1075142621994019, "reward_std": 0.3312613368034363, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6681516170501709, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8826562762260437, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 431.1875, "completions/mean_terminated_length": 402.03277587890625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 1.1408, "grad_norm": 0.054529860615730286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22258726.0, "reward": 1.0597119331359863, "reward_std": 0.422410786151886, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6506718397140503, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8184375166893005, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 446.65625, "completions/mean_terminated_length": 437.4920959472656, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.1416, "grad_norm": 0.04005309194326401, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22302384.0, "reward": 0.803765594959259, "reward_std": 0.24813950061798096, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.4669046998023987, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7326562404632568, "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": 936.0, "completions/max_terminated_length": 936.0, "completions/mean_length": 370.21875, "completions/mean_terminated_length": 370.21875, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.1424, "grad_norm": 0.0440581738948822, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22341758.0, "reward": 1.011565923690796, "reward_std": 0.3644689917564392, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5856297016143799, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8845312595367432, "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": 588.0, "completions/max_terminated_length": 588.0, "completions/mean_length": 362.765625, "completions/mean_terminated_length": 362.765625, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.1432, "grad_norm": 0.05454855412244797, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22386927.0, "reward": 0.728785514831543, "reward_std": 0.35437244176864624, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.2544437646865845, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9681249856948853, "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": 750.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 350.8125, "completions/mean_terminated_length": 350.8125, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.144, "grad_norm": 0.06034263223409653, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22425635.0, "reward": 0.7791368961334229, "reward_std": 0.3682374358177185, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3238968551158905, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9399999976158142, "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": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 438.34375, "completions/mean_terminated_length": 438.34375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.1448, "grad_norm": 0.056306518614292145, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22469177.0, "reward": 1.019457459449768, "reward_std": 0.3162892460823059, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5389124751091003, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9653124809265137, "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": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 441.28125, "completions/mean_terminated_length": 441.28125, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.1456, "grad_norm": 0.056024886667728424, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22515291.0, "reward": 0.9736062288284302, "reward_std": 0.31046611070632935, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.6034609079360962, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7717187404632568, "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": 0.0, "completions/max_length": 696.0, "completions/max_terminated_length": 696.0, "completions/mean_length": 432.640625, "completions/mean_terminated_length": 432.640625, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.1464, "grad_norm": 0.03858473151922226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22559140.0, "reward": 0.8043190240859985, "reward_std": 0.25136393308639526, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.32738593220710754, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.96421879529953, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 533.734375, "completions/mean_terminated_length": 517.9193115234375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 1.1472, "grad_norm": 0.04532468691468239, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22609763.0, "reward": 0.9553883075714111, "reward_std": 0.34763437509536743, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.6607750058174133, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.6465625166893005, "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": 777.0, "completions/max_terminated_length": 777.0, "completions/mean_length": 371.90625, "completions/mean_terminated_length": 371.90625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.148, "grad_norm": 0.0530267208814621, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22650253.0, "reward": 0.6583799123764038, "reward_std": 0.26470836997032166, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.20738281309604645, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9442187547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 503.9375, "completions/mean_terminated_length": 487.1612854003906, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.1488, "grad_norm": 0.06380227208137512, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22698665.0, "reward": 1.0085797309875488, "reward_std": 0.40565264225006104, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.657782793045044, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7264062762260437, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 786.0, "completions/mean_length": 458.640625, "completions/mean_terminated_length": 440.4031982421875, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.1496, "grad_norm": 0.062324825674295425, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22745394.0, "reward": 0.8602274656295776, "reward_std": 0.42200204730033875, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.454828143119812, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8589062690734863, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 472.1875, "completions/mean_terminated_length": 445.0491638183594, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.1504, "grad_norm": 0.047490302473306656, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22791822.0, "reward": 0.9890672564506531, "reward_std": 0.3764258027076721, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5718828439712524, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8520312309265137, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 534.765625, "completions/mean_terminated_length": 493.3050842285156, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 1.1512, "grad_norm": 0.05744870752096176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 22843903.0, "reward": 0.6216400861740112, "reward_std": 0.2962142527103424, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.1964031159877777, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8793749809265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 464.859375, "completions/mean_terminated_length": 455.9841613769531, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 1.152, "grad_norm": 0.049528464674949646, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22889222.0, "reward": 1.129990816116333, "reward_std": 0.36559635400772095, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.7287296056747437, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8285624980926514, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 468.5625, "completions/mean_terminated_length": 459.7460632324219, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.1528, "grad_norm": 0.06270156800746918, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22935738.0, "reward": 1.03458833694458, "reward_std": 0.42501094937324524, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.6473000049591064, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8026562333106995, "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": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 463.4375, "completions/mean_terminated_length": 463.4375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.1536, "grad_norm": 0.044134486466646194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 22980774.0, "reward": 1.230616569519043, "reward_std": 0.38008663058280945, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.7581062316894531, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9343750476837158, "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": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 411.1875, "completions/mean_terminated_length": 411.1875, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 1.1544, "grad_norm": 0.07333852350711823, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23023458.0, "reward": 0.8611165881156921, "reward_std": 0.40628302097320557, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4253562092781067, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8931249380111694, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 540.484375, "completions/mean_terminated_length": 532.8095703125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 1.1552, "grad_norm": 0.04188939929008484, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23075393.0, "reward": 0.8434658050537109, "reward_std": 0.39769190549850464, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.42130470275878906, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8620312809944153, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 514.0, "completions/mean_terminated_length": 514.0, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.156, "grad_norm": 0.04458552226424217, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23124161.0, "reward": 0.851743221282959, "reward_std": 0.36957070231437683, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4066093862056732, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9178125262260437, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 598.8125, "completions/mean_terminated_length": 577.901611328125, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.1568, "grad_norm": 0.0428144671022892, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23179941.0, "reward": 0.9252516031265259, "reward_std": 0.42553988099098206, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.616126537322998, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6473437547683716, "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": 804.0, "completions/max_terminated_length": 804.0, "completions/mean_length": 431.234375, "completions/mean_terminated_length": 431.234375, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.1576, "grad_norm": 0.06810154765844345, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23224452.0, "reward": 0.9509885907173157, "reward_std": 0.4077340066432953, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.48009997606277466, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9596874713897705, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 555.21875, "completions/mean_terminated_length": 532.1638793945312, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.1584, "grad_norm": 0.036334048956632614, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23275762.0, "reward": 0.7161891460418701, "reward_std": 0.29006680846214294, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29175156354904175, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8832812309265137, "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": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 499.09375, "completions/mean_terminated_length": 499.09375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 1.1592, "grad_norm": 0.04006053879857063, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23324168.0, "reward": 0.8794111013412476, "reward_std": 0.25758346915245056, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5244452953338623, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7776562571525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 387.203125, "completions/mean_terminated_length": 377.0952453613281, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.16, "grad_norm": 0.05237425118684769, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23364693.0, "reward": 0.8044798374176025, "reward_std": 0.347247838973999, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.35895779728889465, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9167187213897705, "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": 821.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 467.78125, "completions/mean_terminated_length": 467.78125, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 1.1608, "grad_norm": 0.02461780048906803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23410839.0, "reward": 1.284376621246338, "reward_std": 0.20553870499134064, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.8656265735626221, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8579687476158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 449.1875, "completions/mean_terminated_length": 430.6451416015625, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.1616, "grad_norm": 0.07214079052209854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23454995.0, "reward": 0.9473205804824829, "reward_std": 0.4398452043533325, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4883890748023987, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9245312809944153, "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": 0.0, "completions/max_length": 801.0, "completions/max_terminated_length": 801.0, "completions/mean_length": 521.15625, "completions/mean_terminated_length": 521.15625, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 1.1623999999999999, "grad_norm": 0.04403293505311012, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23504829.0, "reward": 0.8436947464942932, "reward_std": 0.3606581687927246, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37488749623298645, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9524999856948853, "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": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 325.8125, "completions/mean_terminated_length": 325.8125, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 1.1632, "grad_norm": 0.0679711401462555, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23540641.0, "reward": 0.9248479008674622, "reward_std": 0.3426002562046051, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.45906874537467957, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9424999952316284, "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": 0.0, "completions/max_length": 929.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 505.78125, "completions/mean_terminated_length": 505.78125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 1.164, "grad_norm": 0.03442218899726868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23589507.0, "reward": 0.6723822355270386, "reward_std": 0.2643858790397644, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21976250410079956, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9387500286102295, "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": 840.0, "completions/max_terminated_length": 840.0, "completions/mean_length": 468.75, "completions/mean_terminated_length": 468.75, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.1648, "grad_norm": 0.05519532784819603, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23635219.0, "reward": 1.1988728046417236, "reward_std": 0.403117299079895, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7571187019348145, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8982812166213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 530.015625, "completions/mean_terminated_length": 522.1746215820312, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.1656, "grad_norm": 0.05434631556272507, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23684900.0, "reward": 0.9311196804046631, "reward_std": 0.4005400538444519, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5028624534606934, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8823280930519104, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 461.984375, "completions/mean_terminated_length": 414.3559265136719, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.1663999999999999, "grad_norm": 0.06186073273420334, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 23729747.0, "reward": 0.9825807809829712, "reward_std": 0.39833301305770874, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5589093565940857, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8484375476837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 499.078125, "completions/mean_terminated_length": 490.7460632324219, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 1.1672, "grad_norm": 0.04045933857560158, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23780824.0, "reward": 1.0051212310791016, "reward_std": 0.3123239278793335, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5571156740188599, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8893749713897705, "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": 813.0, "completions/max_terminated_length": 813.0, "completions/mean_length": 406.46875, "completions/mean_terminated_length": 406.46875, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 1.168, "grad_norm": 0.056747619062662125, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23822246.0, "reward": 1.0300548076629639, "reward_std": 0.42149245738983154, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5757328271865845, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9189062714576721, "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": 939.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 452.359375, "completions/mean_terminated_length": 452.359375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.1688, "grad_norm": 0.0519891232252121, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23868589.0, "reward": 0.9738079309463501, "reward_std": 0.35506516695022583, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5882390737533569, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8414843678474426, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 534.796875, "completions/mean_terminated_length": 510.7376708984375, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 1.1696, "grad_norm": 0.03292745351791382, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23920896.0, "reward": 0.8665140867233276, "reward_std": 0.36710312962532043, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.5142765045166016, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7504687905311584, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 457.296875, "completions/mean_terminated_length": 448.3016052246094, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.1703999999999999, "grad_norm": 0.04971197620034218, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 23965411.0, "reward": 0.9758096933364868, "reward_std": 0.41785770654678345, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5141172409057617, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9104686975479126, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 498.796875, "completions/mean_terminated_length": 481.8548278808594, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.1712, "grad_norm": 0.046108637005090714, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24013334.0, "reward": 0.8494094610214233, "reward_std": 0.39610961079597473, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4488171935081482, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8335937261581421, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 436.109375, "completions/mean_terminated_length": 436.109375, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.172, "grad_norm": 0.05925459414720535, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24058157.0, "reward": 0.8574008941650391, "reward_std": 0.3811948299407959, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.46480000019073486, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8207812309265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 429.15625, "completions/mean_terminated_length": 419.7143249511719, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.1728, "grad_norm": 0.05507020279765129, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24102695.0, "reward": 0.7790847420692444, "reward_std": 0.32709670066833496, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3237921893596649, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9226562976837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 602.65625, "completions/mean_terminated_length": 595.96826171875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.1736, "grad_norm": 0.04148944839835167, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24157937.0, "reward": 0.7806947231292725, "reward_std": 0.37827759981155396, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3895125091075897, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8221875429153442, "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": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 381.90625, "completions/mean_terminated_length": 381.90625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.1743999999999999, "grad_norm": 0.10258068144321442, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24199163.0, "reward": 1.2448369264602661, "reward_std": 0.3305200934410095, "rewards/accuracy_reward": 0.734375, "rewards/brier_reward": 0.7865468859672546, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9279687404632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 436.34375, "completions/mean_terminated_length": 427.0158996582031, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 1.1752, "grad_norm": 0.09492778033018112, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24245873.0, "reward": 0.8662149906158447, "reward_std": 0.4189029932022095, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4355531334877014, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8917187452316284, "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": 631.0, "completions/max_terminated_length": 631.0, "completions/mean_length": 416.625, "completions/mean_terminated_length": 416.625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.176, "grad_norm": 0.05364704132080078, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24290393.0, "reward": 1.088703989982605, "reward_std": 0.3794032335281372, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.7086561918258667, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7900000214576721, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 477.8125, "completions/mean_terminated_length": 410.7368469238281, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 1.1768, "grad_norm": 0.06689091771841049, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24337165.0, "reward": 0.9613509178161621, "reward_std": 0.4418637752532959, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5789499878883362, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7759374976158142, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 381.953125, "completions/mean_terminated_length": 350.3770446777344, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "epoch": 1.1776, "grad_norm": 0.1120075210928917, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24378266.0, "reward": 0.9532260894775391, "reward_std": 0.489637553691864, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5001999735832214, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.903749942779541, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 548.5, "completions/mean_terminated_length": 460.4444580078125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 1.1784, "grad_norm": 0.04312225431203842, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24431082.0, "reward": 0.85650634765625, "reward_std": 0.3472398519515991, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4786359369754791, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7753125429153442, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 592.109375, "completions/mean_terminated_length": 570.8688354492188, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 1.1792, "grad_norm": 0.03589598461985588, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24484481.0, "reward": 0.7944657802581787, "reward_std": 0.3505416512489319, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.4014297127723694, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8135937452316284, "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": 994.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 384.53125, "completions/mean_terminated_length": 384.53125, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.18, "grad_norm": 0.04171023145318031, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24524819.0, "reward": 1.0248785018920898, "reward_std": 0.32977211475372314, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5341296792030334, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9723437428474426, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 598.125, "completions/mean_terminated_length": 584.3870849609375, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 1.1808, "grad_norm": 0.03543979302048683, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24579243.0, "reward": 0.863682746887207, "reward_std": 0.34552526473999023, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.5711140632629395, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.65296870470047, "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": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 381.953125, "completions/mean_terminated_length": 381.953125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.1816, "grad_norm": 0.06618726253509521, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24619608.0, "reward": 0.9304604530334473, "reward_std": 0.43774986267089844, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4702937602996826, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.93812495470047, "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": 0.0, "completions/max_length": 646.0, "completions/max_terminated_length": 646.0, "completions/mean_length": 367.15625, "completions/mean_terminated_length": 367.15625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 1.1824, "grad_norm": 0.06026383116841316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24658802.0, "reward": 1.118424415588379, "reward_std": 0.3608420491218567, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.643096923828125, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9614062309265137, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 455.828125, "completions/mean_terminated_length": 455.828125, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.1832, "grad_norm": 0.03982439264655113, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24703959.0, "reward": 0.6978815197944641, "reward_std": 0.3175278306007385, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.2707609534263611, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8907812833786011, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 443.546875, "completions/mean_terminated_length": 434.3333435058594, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 1.184, "grad_norm": 0.07047735899686813, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24748298.0, "reward": 1.1060245037078857, "reward_std": 0.47930577397346497, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6495468616485596, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8825000524520874, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 476.5, "completions/mean_terminated_length": 467.8095397949219, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.1848, "grad_norm": 0.04641751945018768, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 24794970.0, "reward": 0.944670557975769, "reward_std": 0.3491273820400238, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.49871405959129333, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9167187213897705, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 479.015625, "completions/mean_terminated_length": 365.9056701660156, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.1856, "grad_norm": 0.04621231183409691, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24844075.0, "reward": 0.651425838470459, "reward_std": 0.3077686131000519, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.27160000801086426, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7928124666213989, "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": 923.0, "completions/max_terminated_length": 923.0, "completions/mean_length": 462.015625, "completions/mean_terminated_length": 462.015625, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.1864, "grad_norm": 0.05862155556678772, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24889932.0, "reward": 1.111506462097168, "reward_std": 0.3130871057510376, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6605109572410583, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9159374833106995, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 473.5, "completions/mean_terminated_length": 455.7419128417969, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.1872, "grad_norm": 0.04690905660390854, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24936300.0, "reward": 0.759638249874115, "reward_std": 0.3275766372680664, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.4099000096321106, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7737499475479126, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 448.21875, "completions/mean_terminated_length": 409.8333435058594, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.188, "grad_norm": 0.06608695536851883, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 24981578.0, "reward": 0.9420400261878967, "reward_std": 0.4362315535545349, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5090781450271606, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8571875095367432, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 691.515625, "completions/mean_terminated_length": 650.6842041015625, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 1.1888, "grad_norm": 0.04289550706744194, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25042731.0, "reward": 0.8347203731536865, "reward_std": 0.42920979857444763, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.46631407737731934, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7429687976837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 373.203125, "completions/mean_terminated_length": 362.873046875, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 1.1896, "grad_norm": 0.07930220663547516, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25081672.0, "reward": 1.2858691215515137, "reward_std": 0.3973260819911957, "rewards/accuracy_reward": 0.796875, "rewards/brier_reward": 0.8061109185218811, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9351562261581421, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 528.015625, "completions/mean_terminated_length": 485.9830627441406, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.1904, "grad_norm": 0.038327861577272415, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25131209.0, "reward": 1.1134507656097412, "reward_std": 0.3706352710723877, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.7112749814987183, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8046875, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 506.609375, "completions/mean_terminated_length": 489.9193420410156, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 1.1912, "grad_norm": 0.023409800603985786, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25180752.0, "reward": 0.7203673720359802, "reward_std": 0.16167396306991577, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.30010780692100525, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8782812356948853, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 523.890625, "completions/mean_terminated_length": 507.758056640625, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 1.192, "grad_norm": 0.035519715398550034, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25232377.0, "reward": 0.8240719437599182, "reward_std": 0.3818798065185547, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.4450171887874603, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7714062929153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 588.46875, "completions/mean_terminated_length": 559.433349609375, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 1.1928, "grad_norm": 0.042274292558431625, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25285671.0, "reward": 1.025586724281311, "reward_std": 0.4170171618461609, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.6605468988418579, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.746874988079071, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 416.4375, "completions/mean_terminated_length": 386.5573425292969, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.1936, "grad_norm": 0.06028430163860321, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25327907.0, "reward": 0.7760884165763855, "reward_std": 0.3336741626262665, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.33342498540878296, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9118750095367432, "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": 746.0, "completions/max_terminated_length": 746.0, "completions/mean_length": 281.71875, "completions/mean_terminated_length": 281.71875, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.1944, "grad_norm": 0.06471502035856247, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25361185.0, "reward": 0.8898558616638184, "reward_std": 0.2850934863090515, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4203343689441681, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9595312476158142, "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": 922.0, "completions/max_terminated_length": 922.0, "completions/mean_length": 339.734375, "completions/mean_terminated_length": 339.734375, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 1.1952, "grad_norm": 0.054599251598119736, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25398720.0, "reward": 0.7175440788269043, "reward_std": 0.2940598726272583, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.2788359522819519, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9017187356948853, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 501.125, "completions/mean_terminated_length": 475.4098205566406, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.196, "grad_norm": 0.04102223366498947, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25447256.0, "reward": 0.6927016377449036, "reward_std": 0.34527960419654846, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.30727654695510864, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8207812309265137, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 519.40625, "completions/mean_terminated_length": 494.59014892578125, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 1.1968, "grad_norm": 0.06338660418987274, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25497522.0, "reward": 0.9498165249824524, "reward_std": 0.43203598260879517, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5715062618255615, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8114062547683716, "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": 0.0, "completions/max_length": 861.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 510.046875, "completions/mean_terminated_length": 510.046875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.1976, "grad_norm": 0.04289114475250244, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25546933.0, "reward": 0.9308761358261108, "reward_std": 0.42195966839790344, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4555000066757202, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9615625143051147, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 438.328125, "completions/mean_terminated_length": 429.0317687988281, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.1984, "grad_norm": 0.05783097818493843, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25591050.0, "reward": 0.9261698126792908, "reward_std": 0.32529503107070923, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4773375988006592, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8868296146392822, "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": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 490.421875, "completions/mean_terminated_length": 490.421875, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.1992, "grad_norm": 0.04943421855568886, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25639621.0, "reward": 1.1177611351013184, "reward_std": 0.3942074179649353, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.7355203628540039, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7967187762260437, "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": 893.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 412.71875, "completions/mean_terminated_length": 412.71875, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.2, "grad_norm": 0.04025683552026749, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25681571.0, "reward": 0.9661151170730591, "reward_std": 0.307282030582428, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.49472811818122864, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9610937237739563, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 428.546875, "completions/mean_terminated_length": 419.09527587890625, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.2008, "grad_norm": 0.05928822606801987, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25725830.0, "reward": 0.7518370747566223, "reward_std": 0.35228851437568665, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.2849218547344208, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9500000476837158, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 604.46875, "completions/mean_terminated_length": 552.9473876953125, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.2016, "grad_norm": 0.04679343104362488, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25781668.0, "reward": 0.7659929990768433, "reward_std": 0.3874339759349823, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.4226093590259552, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7359374761581421, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 619.453125, "completions/mean_terminated_length": 553.2545166015625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 1.2024, "grad_norm": 0.06984096765518188, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25840049.0, "reward": 0.8083375096321106, "reward_std": 0.48022207617759705, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.44479843974113464, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.6942187547683716, "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": 884.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 350.265625, "completions/mean_terminated_length": 350.265625, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.2032, "grad_norm": 0.05080622062087059, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 25879266.0, "reward": 0.851851761341095, "reward_std": 0.2584720253944397, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.5005765557289124, "rewards/confidence_one_or_zero": 0.28125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7235937714576721, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 462.328125, "completions/mean_terminated_length": 462.328125, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.204, "grad_norm": 0.041460830718278885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25923959.0, "reward": 0.7810741066932678, "reward_std": 0.27271342277526855, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.5152719020843506, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.6256250143051147, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 485.0625, "completions/mean_terminated_length": 467.6773986816406, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 1.2048, "grad_norm": 0.042279936373233795, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 25970779.0, "reward": 0.8169658780097961, "reward_std": 0.33221203088760376, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.39955466985702515, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8651562333106995, "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": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 420.53125, "completions/mean_terminated_length": 420.53125, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 1.2056, "grad_norm": 0.051863569766283035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26013021.0, "reward": 0.8479735851287842, "reward_std": 0.2793639898300171, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4146953225135803, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9035937786102295, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 476.375, "completions/mean_terminated_length": 429.9660949707031, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.2064, "grad_norm": 0.013474978506565094, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26060005.0, "reward": 0.9181320667266846, "reward_std": 0.151986226439476, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.6018875241279602, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.6459375023841858, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 523.4375, "completions/mean_terminated_length": 498.81964111328125, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 1.2072, "grad_norm": 0.0530453696846962, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26111153.0, "reward": 0.8313562870025635, "reward_std": 0.44673365354537964, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.42833593487739563, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.844531238079071, "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": 722.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 479.171875, "completions/mean_terminated_length": 479.171875, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 1.208, "grad_norm": 0.05240876227617264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26158588.0, "reward": 1.0052284002304077, "reward_std": 0.37749648094177246, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5417046546936035, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9226561784744263, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 524.625, "completions/mean_terminated_length": 516.6984252929688, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.2088, "grad_norm": 0.060410287231206894, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26208164.0, "reward": 1.0532643795013428, "reward_std": 0.39081892371177673, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5752766132354736, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9182811975479126, "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": 0.0, "completions/max_length": 776.0, "completions/max_terminated_length": 776.0, "completions/mean_length": 414.9375, "completions/mean_terminated_length": 414.9375, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 1.2096, "grad_norm": 0.03906882181763649, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26250992.0, "reward": 1.1608142852783203, "reward_std": 0.2961529493331909, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.7591265439987183, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8421874642372131, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 374.5625, "completions/mean_terminated_length": 364.2539978027344, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 1.2104, "grad_norm": 0.08950353413820267, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26292484.0, "reward": 1.0778744220733643, "reward_std": 0.33190101385116577, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.7026219367980957, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7603124976158142, "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": 872.0, "completions/max_terminated_length": 872.0, "completions/mean_length": 447.8125, "completions/mean_terminated_length": 447.8125, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 1.2112, "grad_norm": 0.048280056565999985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26336568.0, "reward": 0.7241878509521484, "reward_std": 0.2114964723587036, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.2608734369277954, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9585937261581421, "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": 0.0, "completions/max_length": 834.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 335.34375, "completions/mean_terminated_length": 335.34375, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "epoch": 1.212, "grad_norm": 0.07106524705886841, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26374014.0, "reward": 0.9514402747154236, "reward_std": 0.46877267956733704, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.48100313544273376, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9365625381469727, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 507.65625, "completions/mean_terminated_length": 499.4603576660156, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 1.2128, "grad_norm": 0.052640728652477264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26422440.0, "reward": 0.9165806770324707, "reward_std": 0.32104581594467163, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4894093871116638, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8787499666213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 627.421875, "completions/mean_terminated_length": 621.1270141601562, "completions/min_length": 311.0, "completions/min_terminated_length": 311.0, "epoch": 1.2136, "grad_norm": 0.025291744619607925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26478643.0, "reward": 0.7187618613243103, "reward_std": 0.2913065552711487, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.35939687490463257, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7903124690055847, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 470.375, "completions/mean_terminated_length": 452.51611328125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 1.2144, "grad_norm": 0.05014880374073982, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26525755.0, "reward": 0.8810641765594482, "reward_std": 0.26815295219421387, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.5433765649795532, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7039062976837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 824.0, "completions/mean_length": 458.640625, "completions/mean_terminated_length": 449.66668701171875, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.2152, "grad_norm": 0.03263864666223526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26571172.0, "reward": 0.6055673360824585, "reward_std": 0.21757429838180542, "rewards/accuracy_reward": 0.046875, "rewards/brier_reward": 0.19550782442092896, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8929687738418579, "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": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 528.46875, "completions/mean_terminated_length": 528.46875, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.216, "grad_norm": 0.03894927352666855, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26621474.0, "reward": 0.8542150855064392, "reward_std": 0.27964773774147034, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39592814445495605, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9418750405311584, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 451.0, "completions/mean_terminated_length": 441.90478515625, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.2168, "grad_norm": 0.04840056598186493, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26666450.0, "reward": 1.255171298980713, "reward_std": 0.34152525663375854, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.8853405714035034, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7332812547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 475.203125, "completions/mean_terminated_length": 457.5, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.2176, "grad_norm": 0.054615359753370285, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26712367.0, "reward": 0.9209555983543396, "reward_std": 0.3846433758735657, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.5137843489646912, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.84375, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 606.921875, "completions/mean_terminated_length": 586.4097900390625, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 1.2184, "grad_norm": 0.03406061977148056, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26768218.0, "reward": 0.8037735223770142, "reward_std": 0.370273232460022, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.40442031621932983, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8617187738418579, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 567.4375, "completions/mean_terminated_length": 502.21429443359375, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.2192, "grad_norm": 0.08056768029928207, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26820502.0, "reward": 0.7941008806228638, "reward_std": 0.44181784987449646, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4475749731063843, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7118749618530273, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 551.421875, "completions/mean_terminated_length": 502.53448486328125, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.22, "grad_norm": 0.041064538061618805, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26871793.0, "reward": 0.7910827994346619, "reward_std": 0.4413678050041199, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.45716404914855957, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7032812833786011, "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": 0.0, "completions/max_length": 683.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 383.09375, "completions/mean_terminated_length": 383.09375, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.2208, "grad_norm": 0.062457937747240067, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26912455.0, "reward": 0.907568097114563, "reward_std": 0.34858521819114685, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4401344060897827, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9471875429153442, "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": 651.0, "completions/max_terminated_length": 651.0, "completions/mean_length": 323.84375, "completions/mean_terminated_length": 323.84375, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 1.2216, "grad_norm": 0.05908520892262459, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 26950541.0, "reward": 1.0100321769714355, "reward_std": 0.3019281327724457, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5356875061988831, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 492.25, "completions/mean_terminated_length": 475.0967712402344, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.2224, "grad_norm": 0.05278084799647331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 26999613.0, "reward": 0.832314133644104, "reward_std": 0.31018882989883423, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.4927515983581543, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7729687690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 517.0625, "completions/mean_terminated_length": 509.0158996582031, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 1.2232, "grad_norm": 0.0451519675552845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27050097.0, "reward": 0.8383354544639587, "reward_std": 0.38685011863708496, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4110437333583832, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8862500190734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 880.0, "completions/mean_length": 485.6875, "completions/mean_terminated_length": 477.14288330078125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.224, "grad_norm": 0.035379454493522644, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27098045.0, "reward": 0.7960219979286194, "reward_std": 0.27341461181640625, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.4357921779155731, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7629531621932983, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 644.875, "completions/mean_terminated_length": 598.3157958984375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 1.2248, "grad_norm": 0.02779213711619377, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27156837.0, "reward": 0.6742389798164368, "reward_std": 0.3009292781352997, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.3641015291213989, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7164062261581421, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 463.1875, "completions/mean_terminated_length": 435.6065368652344, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.2256, "grad_norm": 0.05411571264266968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27202369.0, "reward": 0.7893697023391724, "reward_std": 0.2985537648200989, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3756124973297119, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8556250333786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 487.6875, "completions/mean_terminated_length": 451.933349609375, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.2264, "grad_norm": 0.05549689382314682, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27248813.0, "reward": 0.6428635120391846, "reward_std": 0.3491060733795166, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.23884999752044678, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8504687547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 890.0, "completions/mean_length": 580.828125, "completions/mean_terminated_length": 566.5322265625, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 1.2272, "grad_norm": 0.03369314968585968, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27303810.0, "reward": 0.7577986121177673, "reward_std": 0.33213353157043457, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.32809531688690186, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8935937881469727, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 624.59375, "completions/mean_terminated_length": 559.236328125, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 1.228, "grad_norm": 0.032611098140478134, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27361608.0, "reward": 0.8528133630752563, "reward_std": 0.3474388122558594, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5181249976158142, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.702343761920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 508.53125, "completions/mean_terminated_length": 500.3492431640625, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 1.2288000000000001, "grad_norm": 0.049349479377269745, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27410106.0, "reward": 0.8741540908813477, "reward_std": 0.43886229395866394, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4358062446117401, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8996875286102295, "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": 0.203125, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 678.609375, "completions/mean_terminated_length": 590.5686645507812, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.2296, "grad_norm": 0.0628722682595253, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27470097.0, "reward": 0.7440000176429749, "reward_std": 0.4938725233078003, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.37862342596054077, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.78125, "rewards/mean_confidence_reward": 0.7540624737739563, "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": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 437.015625, "completions/mean_terminated_length": 437.015625, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.2304, "grad_norm": 0.045703429728746414, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27513954.0, "reward": 0.9010009765625, "reward_std": 0.2860170006752014, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4270000159740448, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9587500095367432, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 425.59375, "completions/mean_terminated_length": 406.2903137207031, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.2312, "grad_norm": 0.08210164308547974, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27556712.0, "reward": 0.9009643197059631, "reward_std": 0.45732858777046204, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.44255155324935913, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9217187166213989, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 536.53125, "completions/mean_terminated_length": 486.10345458984375, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 1.232, "grad_norm": 0.048792764544487, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27606506.0, "reward": 1.0706603527069092, "reward_std": 0.3876461386680603, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6569437384605408, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8284375071525574, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 597.84375, "completions/mean_terminated_length": 553.7586059570312, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 1.2328000000000001, "grad_norm": 0.056855130940675735, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27662144.0, "reward": 1.077366590499878, "reward_std": 0.44172203540802, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.670356273651123, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8100000023841858, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 607.734375, "completions/mean_terminated_length": 539.6181640625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 1.2336, "grad_norm": 0.03200100362300873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27717471.0, "reward": 0.9981062412261963, "reward_std": 0.2662375867366791, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6837109327316284, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.640625, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 586.1875, "completions/mean_terminated_length": 572.0645141601562, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 1.2344, "grad_norm": 0.046086739748716354, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27771579.0, "reward": 0.9328721761703491, "reward_std": 0.49712616205215454, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.49074220657348633, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9128124713897705, "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": 0.0, "completions/max_length": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 412.953125, "completions/mean_terminated_length": 412.953125, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 1.2352, "grad_norm": 0.04566653445363045, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 27816056.0, "reward": 0.8544508218765259, "reward_std": 0.3107070028781891, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.505774974822998, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7746875286102295, "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": 0.234375, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 704.03125, "completions/mean_terminated_length": 606.0816040039062, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.236, "grad_norm": 0.060081783682107925, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27877530.0, "reward": 0.780642032623291, "reward_std": 0.5287032127380371, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.49878281354904175, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.765625, "rewards/mean_confidence_reward": 0.5960937738418579, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 481.703125, "completions/mean_terminated_length": 435.7457580566406, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.2368000000000001, "grad_norm": 0.05359148979187012, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27924855.0, "reward": 0.9230282306671143, "reward_std": 0.39683622121810913, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5023046731948853, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8582812547683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 483.3125, "completions/mean_terminated_length": 474.7301940917969, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.2376, "grad_norm": 0.07004538178443909, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 27971883.0, "reward": 0.9169306755065918, "reward_std": 0.43162623047828674, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.490109384059906, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8953125476837158, "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": 0.0, "completions/max_length": 991.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 492.265625, "completions/mean_terminated_length": 492.265625, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.2384, "grad_norm": 0.06603068858385086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28019388.0, "reward": 0.8145244121551514, "reward_std": 0.33406946063041687, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.36342185735702515, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9331250190734863, "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": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 456.515625, "completions/mean_terminated_length": 456.515625, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 1.2392, "grad_norm": 0.04718991741538048, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28065245.0, "reward": 0.8700103759765625, "reward_std": 0.31224262714385986, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.458768755197525, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8759375214576721, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 558.265625, "completions/mean_terminated_length": 543.241943359375, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 1.24, "grad_norm": 0.0475226566195488, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28119582.0, "reward": 0.7196882963180542, "reward_std": 0.335529625415802, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.36124998331069946, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7874999642372131, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 434.15625, "completions/mean_terminated_length": 415.1290283203125, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 1.2408, "grad_norm": 0.07249294221401215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28163928.0, "reward": 0.7038079500198364, "reward_std": 0.367534875869751, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.3138640522956848, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8114062547683716, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 615.765625, "completions/mean_terminated_length": 573.5344848632812, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 1.2416, "grad_norm": 0.0456414557993412, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28221289.0, "reward": 0.7903704047203064, "reward_std": 0.4282055199146271, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.4713640809059143, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7035937309265137, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 562.796875, "completions/mean_terminated_length": 487.3272705078125, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 1.2424, "grad_norm": 0.05753902345895767, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28274620.0, "reward": 0.7753665447235107, "reward_std": 0.4579416513442993, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.37885624170303345, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.8156249523162842, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 844.0, "completions/mean_length": 456.25, "completions/mean_terminated_length": 437.93548583984375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.2432, "grad_norm": 0.03819156065583229, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28320188.0, "reward": 0.7404862642288208, "reward_std": 0.26996612548828125, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.2934703230857849, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9139062166213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 418.609375, "completions/mean_terminated_length": 409.0000305175781, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.244, "grad_norm": 0.05211174488067627, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28362723.0, "reward": 0.8875908255577087, "reward_std": 0.37650370597839355, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4314297139644623, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9314062595367432, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 434.890625, "completions/mean_terminated_length": 362.5438537597656, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 1.2448, "grad_norm": 0.08078719675540924, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28406732.0, "reward": 0.8895642757415771, "reward_std": 0.35481950640678406, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4666265845298767, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8548437356948853, "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": 673.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 377.609375, "completions/mean_terminated_length": 377.609375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 1.2456, "grad_norm": 0.07581423223018646, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28445763.0, "reward": 1.1458477973937988, "reward_std": 0.4211430549621582, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.7291937470436096, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8368749618530273, "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": 929.0, "completions/max_terminated_length": 929.0, "completions/mean_length": 409.984375, "completions/mean_terminated_length": 409.984375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 1.2464, "grad_norm": 0.043810341507196426, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28487170.0, "reward": 1.0540345907211304, "reward_std": 0.28131335973739624, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6393171548843384, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8656250238418579, "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": 0.0, "completions/max_length": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 464.609375, "completions/mean_terminated_length": 464.609375, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.2472, "grad_norm": 0.05380271002650261, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28532025.0, "reward": 0.9720714092254639, "reward_std": 0.3571241497993469, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.537890613079071, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8946718573570251, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 457.203125, "completions/mean_terminated_length": 448.20635986328125, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "epoch": 1.248, "grad_norm": 0.07053760439157486, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28577382.0, "reward": 1.0662860870361328, "reward_std": 0.3805201053619385, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6013203263282776, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9276562333106995, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 545.109375, "completions/mean_terminated_length": 513.183349609375, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 1.2488, "grad_norm": 0.06221984326839447, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28627549.0, "reward": 1.1365963220596313, "reward_std": 0.46156132221221924, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6950656175613403, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8756250143051147, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 465.609375, "completions/mean_terminated_length": 456.7460632324219, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 1.2496, "grad_norm": 0.04596773162484169, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28672852.0, "reward": 0.8041220903396606, "reward_std": 0.3811837434768677, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3738671541213989, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8557640314102173, "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": 806.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 472.28125, "completions/mean_terminated_length": 472.28125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.2504, "grad_norm": 0.035877835005521774, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28721446.0, "reward": 1.106579303741455, "reward_std": 0.3398491144180298, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6350312232971191, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 457.78125, "completions/mean_terminated_length": 439.51611328125, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 1.2511999999999999, "grad_norm": 0.05643845349550247, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 28766680.0, "reward": 1.0443267822265625, "reward_std": 0.41286584734916687, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6199015378952026, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8473437428474426, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 478.5, "completions/mean_terminated_length": 451.672119140625, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.252, "grad_norm": 0.056913111358881, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28813432.0, "reward": 1.0467369556427002, "reward_std": 0.36878910660743713, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.593471884727478, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8971874713897705, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 468.671875, "completions/mean_terminated_length": 431.6500244140625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.2528000000000001, "grad_norm": 0.03346291556954384, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28860035.0, "reward": 0.7339391708374023, "reward_std": 0.2681478261947632, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3428765535354614, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.82421875, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 634.046875, "completions/mean_terminated_length": 593.7069091796875, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.2536, "grad_norm": 0.04203236848115921, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28917942.0, "reward": 0.7179765105247498, "reward_std": 0.3620987832546234, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.4047015607357025, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7182812690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 951.0, "completions/mean_length": 497.6875, "completions/mean_terminated_length": 489.3333740234375, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 1.2544, "grad_norm": 0.03744041174650192, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 28966338.0, "reward": 0.8546908497810364, "reward_std": 0.32323312759399414, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4437547028064728, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8653125166893005, "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": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 532.96875, "completions/mean_terminated_length": 532.96875, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 1.2551999999999999, "grad_norm": 0.05324840545654297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29016976.0, "reward": 0.5871985554695129, "reward_std": 0.2624598443508148, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.20564529299736023, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8935937881469727, "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": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 333.109375, "completions/mean_terminated_length": 333.109375, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 1.256, "grad_norm": 0.06338287889957428, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29054743.0, "reward": 0.9366994500160217, "reward_std": 0.36477571725845337, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4671468734741211, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9614046812057495, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 448.734375, "completions/mean_terminated_length": 439.60321044921875, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.2568, "grad_norm": 0.053076282143592834, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29097974.0, "reward": 0.7968830466270447, "reward_std": 0.293457567691803, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.390639066696167, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8054687976837158, "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": 976.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 478.265625, "completions/mean_terminated_length": 478.265625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.2576, "grad_norm": 0.056155089288949966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29145511.0, "reward": 0.9289213418960571, "reward_std": 0.3683028817176819, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4984656274318695, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8812500238418579, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 551.234375, "completions/mean_terminated_length": 527.9835815429688, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.2584, "grad_norm": 0.03848298266530037, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29197798.0, "reward": 0.9050737619400024, "reward_std": 0.32036155462265015, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4663953185081482, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8882812261581421, "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": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 441.875, "completions/mean_terminated_length": 441.875, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.2591999999999999, "grad_norm": 0.0555666908621788, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29242254.0, "reward": 0.8801507949829102, "reward_std": 0.32973501086235046, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.49467501044273376, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7681249976158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 405.5, "completions/mean_terminated_length": 385.5483703613281, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.26, "grad_norm": 0.031101442873477936, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29285630.0, "reward": 0.9855509400367737, "reward_std": 0.1912582963705063, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5492249727249146, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8918750286102295, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 507.984375, "completions/mean_terminated_length": 464.2542419433594, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 1.2608, "grad_norm": 0.06669484823942184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29335181.0, "reward": 0.9798359870910645, "reward_std": 0.45251914858818054, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.6159203052520752, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7520312666893005, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 806.0, "completions/mean_length": 438.640625, "completions/mean_terminated_length": 409.8524475097656, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 1.2616, "grad_norm": 0.0825200155377388, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29378214.0, "reward": 0.9276391267776489, "reward_std": 0.42130500078201294, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5115265250205994, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8535781502723694, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 517.453125, "completions/mean_terminated_length": 474.5254211425781, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.2624, "grad_norm": 0.06040314957499504, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29427539.0, "reward": 1.0008883476257324, "reward_std": 0.4655555486679077, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.6267750263214111, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7687499523162842, "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": 614.0, "completions/max_terminated_length": 614.0, "completions/mean_length": 343.015625, "completions/mean_terminated_length": 343.015625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.2631999999999999, "grad_norm": 0.04646599665284157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29466244.0, "reward": 1.3590072393417358, "reward_std": 0.2669958770275116, "rewards/accuracy_reward": 0.84375, "rewards/brier_reward": 0.8742625117301941, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462500214576721, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 692.171875, "completions/mean_terminated_length": 675.8524169921875, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 1.264, "grad_norm": 0.04380285367369652, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29528207.0, "reward": 0.8162047863006592, "reward_std": 0.39563533663749695, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4449077844619751, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.780468761920929, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 520.46875, "completions/mean_terminated_length": 504.2257995605469, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.2648, "grad_norm": 0.04958205670118332, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29577277.0, "reward": 0.7765719890594482, "reward_std": 0.40981006622314453, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.36564216017723083, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8743593692779541, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 504.4375, "completions/mean_terminated_length": 478.8852233886719, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.2656, "grad_norm": 0.058824680745601654, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29626425.0, "reward": 0.8198572397232056, "reward_std": 0.3169676959514618, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.38971251249313354, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.885937511920929, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 651.953125, "completions/mean_terminated_length": 598.8035888671875, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 1.2664, "grad_norm": 0.039174217730760574, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29684598.0, "reward": 0.7147234082221985, "reward_std": 0.3404347002506256, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.4450703263282776, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.5854687690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 574.4375, "completions/mean_terminated_length": 567.3016357421875, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 1.2671999999999999, "grad_norm": 0.04272704944014549, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29737618.0, "reward": 0.8257737159729004, "reward_std": 0.36043688654899597, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.38592028617858887, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8892030715942383, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 767.0, "completions/mean_length": 445.8125, "completions/mean_terminated_length": 427.1612854003906, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 1.268, "grad_norm": 0.05588303878903389, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29782806.0, "reward": 0.767821192741394, "reward_std": 0.34100478887557983, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3950155973434448, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7888906002044678, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 470.765625, "completions/mean_terminated_length": 461.9841613769531, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.2688, "grad_norm": 0.0682910680770874, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29827543.0, "reward": 1.0985400676727295, "reward_std": 0.45756518840789795, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6345781683921814, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9262499809265137, "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": 894.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 465.546875, "completions/mean_terminated_length": 465.546875, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.2696, "grad_norm": 0.08344040811061859, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29874026.0, "reward": 0.9882743954658508, "reward_std": 0.4119247794151306, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5702968835830688, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8443750143051147, "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": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 480.171875, "completions/mean_terminated_length": 480.171875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 1.2704, "grad_norm": 0.06150594353675842, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 29922693.0, "reward": 0.7756049036979675, "reward_std": 0.3476894795894623, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.3637077808380127, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8887499570846558, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 588.625, "completions/mean_terminated_length": 543.586181640625, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 1.2711999999999999, "grad_norm": 0.03327392041683197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 29979117.0, "reward": 0.6945118308067322, "reward_std": 0.40064865350723267, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.3733968734741211, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7014062404632568, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 447.3125, "completions/mean_terminated_length": 398.440673828125, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.272, "grad_norm": 0.06747777760028839, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30023985.0, "reward": 1.0120376348495483, "reward_std": 0.4528874158859253, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6021984815597534, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.82421875, "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": 812.0, "completions/max_terminated_length": 812.0, "completions/mean_length": 416.828125, "completions/mean_terminated_length": 416.828125, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 1.2728, "grad_norm": 0.055278897285461426, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30066022.0, "reward": 0.858450174331665, "reward_std": 0.38721609115600586, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.40439844131469727, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8732812404632568, "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": 0.0, "completions/max_length": 522.0, "completions/max_terminated_length": 522.0, "completions/mean_length": 291.890625, "completions/mean_terminated_length": 291.890625, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 1.2736, "grad_norm": 0.06925549358129501, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30099615.0, "reward": 1.2711236476898193, "reward_std": 0.35371071100234985, "rewards/accuracy_reward": 0.734375, "rewards/brier_reward": 0.8078703284263611, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9092187881469727, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 475.03125, "completions/mean_terminated_length": 457.32257080078125, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.2744, "grad_norm": 0.03299914672970772, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30145537.0, "reward": 1.0599009990692139, "reward_std": 0.265680193901062, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6354249715805054, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8782812356948853, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 653.0, "completions/mean_terminated_length": 621.559326171875, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 1.2752, "grad_norm": 0.04493138566613197, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30204369.0, "reward": 0.8775633573532104, "reward_std": 0.4228721857070923, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4738749861717224, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.83984375, "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": 0.265625, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 601.234375, "completions/mean_terminated_length": 448.3191223144531, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 1.276, "grad_norm": 0.06451296806335449, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30260112.0, "reward": 0.7068890333175659, "reward_std": 0.32601937651634216, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.38252657651901245, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.734375, "rewards/mean_confidence_reward": 0.5595312118530273, "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": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 473.15625, "completions/mean_terminated_length": 473.15625, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.2768, "grad_norm": 0.03177356719970703, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30306778.0, "reward": 1.1846369504928589, "reward_std": 0.34307461977005005, "rewards/accuracy_reward": 0.671875, "rewards/brier_reward": 0.6973968744277954, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9600000381469727, "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": 810.0, "completions/max_terminated_length": 810.0, "completions/mean_length": 392.890625, "completions/mean_terminated_length": 392.890625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.2776, "grad_norm": 0.07471024990081787, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30347299.0, "reward": 0.9229604005813599, "reward_std": 0.42673060297966003, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.455293744802475, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9456250071525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 899.0, "completions/mean_length": 460.265625, "completions/mean_terminated_length": 451.3174743652344, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.2784, "grad_norm": 0.06255394220352173, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30393876.0, "reward": 1.1033204793930054, "reward_std": 0.34564924240112305, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.7066390514373779, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8064062595367432, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 579.796875, "completions/mean_terminated_length": 525.24560546875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.2792, "grad_norm": 0.047538794577121735, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30451687.0, "reward": 0.784229040145874, "reward_std": 0.3623906970024109, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.42783123254776, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7509374618530273, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 551.578125, "completions/mean_terminated_length": 544.0794067382812, "completions/min_length": 382.0, "completions/min_terminated_length": 382.0, "epoch": 1.28, "grad_norm": 0.04933956265449524, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30502828.0, "reward": 0.9943314790725708, "reward_std": 0.4479231536388397, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5355359315872192, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9207812547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 430.65625, "completions/mean_terminated_length": 411.51611328125, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.2808, "grad_norm": 0.04547116160392761, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30546182.0, "reward": 0.7119118571281433, "reward_std": 0.28011953830718994, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.2988218665122986, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8743749856948853, "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": 970.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 374.203125, "completions/mean_terminated_length": 374.203125, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "epoch": 1.2816, "grad_norm": 0.04957852140069008, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30587443.0, "reward": 1.2015267610549927, "reward_std": 0.21458441019058228, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.8093016147613525, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8023437857627869, "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": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 452.671875, "completions/mean_terminated_length": 452.671875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.2824, "grad_norm": 0.054172318428754807, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30635182.0, "reward": 0.9793461561203003, "reward_std": 0.3316839933395386, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5680655837059021, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8028124570846558, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 500.875, "completions/mean_terminated_length": 484.0, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 1.2832, "grad_norm": 0.06119895726442337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30683094.0, "reward": 1.0796587467193604, "reward_std": 0.3748689889907837, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6593155860900879, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8631249666213989, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 493.28125, "completions/mean_terminated_length": 438.3793029785156, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.284, "grad_norm": 0.05253171548247337, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30730552.0, "reward": 0.847075879573822, "reward_std": 0.3121541142463684, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.42852500081062317, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8521875143051147, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 584.84375, "completions/mean_terminated_length": 530.9122924804688, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.2848, "grad_norm": 0.040891654789447784, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30784750.0, "reward": 0.9011257886886597, "reward_std": 0.2802080512046814, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.6147500276565552, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6218750476837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 641.0, "completions/mean_length": 434.515625, "completions/mean_terminated_length": 425.15875244140625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.2856, "grad_norm": 0.05694107711315155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30828111.0, "reward": 1.090185284614563, "reward_std": 0.41599440574645996, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6178687810897827, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9378125071525574, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 499.25, "completions/mean_terminated_length": 464.2666931152344, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.2864, "grad_norm": 0.049438346177339554, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 30877679.0, "reward": 0.7508751153945923, "reward_std": 0.32674163579940796, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.3767484426498413, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7993749976158142, "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": 611.0, "completions/max_terminated_length": 611.0, "completions/mean_length": 289.484375, "completions/mean_terminated_length": 289.484375, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.2872, "grad_norm": 0.06460945308208466, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30911966.0, "reward": 1.046494722366333, "reward_std": 0.21944010257720947, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5617375373840332, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9584218859672546, "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": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 431.109375, "completions/mean_terminated_length": 431.109375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.288, "grad_norm": 0.04294288903474808, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 30956837.0, "reward": 1.0375189781188965, "reward_std": 0.28917843103408813, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5594109296798706, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8957812786102295, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 543.328125, "completions/mean_terminated_length": 527.8225708007812, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 1.2888, "grad_norm": 0.02720510959625244, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31007914.0, "reward": 1.0606088638305664, "reward_std": 0.28079622983932495, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6368405818939209, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8665624856948853, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 526.09375, "completions/mean_terminated_length": 510.0322570800781, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 1.2896, "grad_norm": 0.043499574065208435, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31059520.0, "reward": 0.8225884437561035, "reward_std": 0.29693615436553955, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.4107999801635742, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8528124690055847, "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": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 429.3125, "completions/mean_terminated_length": 429.3125, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 1.2904, "grad_norm": 0.06830660253763199, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31103476.0, "reward": 0.9843634366989136, "reward_std": 0.24306908249855042, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5624749660491943, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8621875047683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 392.703125, "completions/mean_terminated_length": 382.68255615234375, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 1.2912, "grad_norm": 0.06185173615813255, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31145137.0, "reward": 0.9992157220840454, "reward_std": 0.42520418763160706, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.6390547156333923, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7560937404632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 408.578125, "completions/mean_terminated_length": 398.8095397949219, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.292, "grad_norm": 0.06139004975557327, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31187366.0, "reward": 1.0761643648147583, "reward_std": 0.4194018840789795, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.5898265838623047, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9142187833786011, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 571.3125, "completions/mean_terminated_length": 549.0491333007812, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.2928, "grad_norm": 0.04873720183968544, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31239786.0, "reward": 0.7169680595397949, "reward_std": 0.32365280389785767, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.34018436074256897, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8031249642372131, "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": 875.0, "completions/max_terminated_length": 875.0, "completions/mean_length": 429.765625, "completions/mean_terminated_length": 429.765625, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.2936, "grad_norm": 0.07632818073034286, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31283563.0, "reward": 0.8302056789398193, "reward_std": 0.2878199517726898, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.4729093909263611, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7895312309265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 504.78125, "completions/mean_terminated_length": 488.0322570800781, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.2944, "grad_norm": 0.06111682578921318, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31334829.0, "reward": 0.7033110857009888, "reward_std": 0.27525562047958374, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.3284953236579895, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7942187190055847, "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": 498.0, "completions/max_terminated_length": 498.0, "completions/mean_length": 288.53125, "completions/mean_terminated_length": 288.53125, "completions/min_length": 104.0, "completions/min_terminated_length": 104.0, "epoch": 1.2952, "grad_norm": 0.07671785354614258, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31367791.0, "reward": 1.1749329566955566, "reward_std": 0.17190513014793396, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6936140656471252, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9567188024520874, "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": 958.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 521.859375, "completions/mean_terminated_length": 521.859375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.296, "grad_norm": 0.025130564346909523, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31418150.0, "reward": 0.7186524868011475, "reward_std": 0.20701727271080017, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.3591781258583069, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7631250023841858, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 503.71875, "completions/mean_terminated_length": 449.89654541015625, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.2968, "grad_norm": 0.08945115655660629, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31466244.0, "reward": 0.97779381275177, "reward_std": 0.5012706518173218, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6118359565734863, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7595312595367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 433.46875, "completions/mean_terminated_length": 394.1000061035156, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 1.2976, "grad_norm": 0.06222410127520561, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31510066.0, "reward": 1.1018362045288086, "reward_std": 0.39248496294021606, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.67242032289505, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8454687595367432, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 542.3125, "completions/mean_terminated_length": 492.4827575683594, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.2984, "grad_norm": 0.06547354906797409, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31561286.0, "reward": 0.9522852301597595, "reward_std": 0.46292734146118164, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5451937913894653, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7709375023841858, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 416.328125, "completions/mean_terminated_length": 406.68255615234375, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 1.2992, "grad_norm": 0.053551845252513885, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31603931.0, "reward": 0.7362784147262573, "reward_std": 0.3453342318534851, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.28505468368530273, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9314062595367432, "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": 987.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 458.59375, "completions/mean_terminated_length": 458.59375, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.3, "grad_norm": 0.064813993871212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31651697.0, "reward": 0.9732891917228699, "reward_std": 0.4345937669277191, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4934515357017517, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9354687929153442, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 525.1875, "completions/mean_terminated_length": 509.0967712402344, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.3008, "grad_norm": 0.04770620912313461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31701101.0, "reward": 1.0139524936676025, "reward_std": 0.3013019859790802, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.6529030799865723, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7321875095367432, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 485.65625, "completions/mean_terminated_length": 477.11114501953125, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.3016, "grad_norm": 0.022442501038312912, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31750391.0, "reward": 0.7182954549789429, "reward_std": 0.2114134430885315, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.3428390622138977, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7963906526565552, "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": 0.0, "completions/max_length": 790.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 439.59375, "completions/mean_terminated_length": 439.59375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.3024, "grad_norm": 0.09239853918552399, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31794557.0, "reward": 0.979364275932312, "reward_std": 0.4409602880477905, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5212265253067017, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9173437356948853, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 534.53125, "completions/mean_terminated_length": 483.89654541015625, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 1.3032, "grad_norm": 0.06621894985437393, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31845119.0, "reward": 0.8561533093452454, "reward_std": 0.4381236135959625, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4466796815395355, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8489062190055847, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 537.796875, "completions/mean_terminated_length": 496.5932312011719, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 1.304, "grad_norm": 0.08722609281539917, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 31895074.0, "reward": 0.9016915559768677, "reward_std": 0.4527190327644348, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5065062642097473, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8153125047683716, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 688.21875, "completions/mean_terminated_length": 653.4827270507812, "completions/min_length": 324.0, "completions/min_terminated_length": 324.0, "epoch": 1.3048, "grad_norm": 0.04773704707622528, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 31955712.0, "reward": 0.6936492919921875, "reward_std": 0.4080084264278412, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3247968554496765, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7979687452316284, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 713.015625, "completions/mean_terminated_length": 662.1272583007812, "completions/min_length": 314.0, "completions/min_terminated_length": 314.0, "epoch": 1.3056, "grad_norm": 0.03134338557720184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32018673.0, "reward": 0.672763466835022, "reward_std": 0.4321001172065735, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.26739999651908875, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8218749761581421, "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": 0.0, "completions/max_length": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 354.375, "completions/mean_terminated_length": 354.375, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.3064, "grad_norm": 0.07221519947052002, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32057273.0, "reward": 0.9233769178390503, "reward_std": 0.34214675426483154, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4405015707015991, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9701562523841858, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 468.0, "completions/mean_terminated_length": 440.65570068359375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 1.3072, "grad_norm": 0.03115226700901985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32104841.0, "reward": 0.8253862261772156, "reward_std": 0.18108168244361877, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.41639530658721924, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8504687547683716, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 563.8125, "completions/mean_terminated_length": 516.2069091796875, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.308, "grad_norm": 0.0418444499373436, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32157549.0, "reward": 1.01313054561615, "reward_std": 0.31376010179519653, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6356344223022461, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8003125190734863, "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": 833.0, "completions/max_terminated_length": 833.0, "completions/mean_length": 394.59375, "completions/mean_terminated_length": 394.59375, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.3088, "grad_norm": 0.05087652429938316, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32200483.0, "reward": 1.0950658321380615, "reward_std": 0.18202444911003113, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6432546973228455, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8892187476158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 470.203125, "completions/mean_terminated_length": 452.33868408203125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.3096, "grad_norm": 0.045378074049949646, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32246464.0, "reward": 1.2187952995300293, "reward_std": 0.2910972535610199, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.8282140493392944, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.80078125, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 597.828125, "completions/mean_terminated_length": 591.0635375976562, "completions/min_length": 351.0, "completions/min_terminated_length": 351.0, "epoch": 1.3104, "grad_norm": 0.04949896037578583, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32301237.0, "reward": 0.8687845468521118, "reward_std": 0.47690820693969727, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4406921863555908, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8592187166213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 476.609375, "completions/mean_terminated_length": 458.95159912109375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 1.3112, "grad_norm": 0.04028324410319328, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32348540.0, "reward": 0.8510180711746216, "reward_std": 0.19782571494579315, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4051593840122223, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8893749713897705, "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": 827.0, "completions/max_terminated_length": 827.0, "completions/mean_length": 353.546875, "completions/mean_terminated_length": 353.546875, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 1.312, "grad_norm": 0.08846881985664368, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32386511.0, "reward": 1.1051580905914307, "reward_std": 0.3502926826477051, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.7259390354156494, "rewards/confidence_one_or_zero": 0.28125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7260937690734863, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 495.59375, "completions/mean_terminated_length": 460.36669921875, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.3128, "grad_norm": 0.06257642060518265, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32433589.0, "reward": 0.7906383872032166, "reward_std": 0.3315812647342682, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.36252498626708984, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8656250238418579, "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": 753.0, "completions/max_terminated_length": 753.0, "completions/mean_length": 428.59375, "completions/mean_terminated_length": 428.59375, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 1.3136, "grad_norm": 0.04509107768535614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32478875.0, "reward": 0.8429072499275208, "reward_std": 0.3056159019470215, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.43581247329711914, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8531249761581421, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 607.34375, "completions/mean_terminated_length": 530.1851806640625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.3144, "grad_norm": 0.03695479407906532, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32533249.0, "reward": 0.5933899283409119, "reward_std": 0.35254067182540894, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21802812814712524, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.7784374952316284, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 610.03125, "completions/mean_terminated_length": 589.672119140625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.3152, "grad_norm": 0.04020287096500397, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32589795.0, "reward": 0.7139643430709839, "reward_std": 0.3848656415939331, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.31855154037475586, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8387500047683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 485.3125, "completions/mean_terminated_length": 476.7619323730469, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 1.316, "grad_norm": 0.04248309135437012, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32636279.0, "reward": 1.0537869930267334, "reward_std": 0.29785606265068054, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6231968402862549, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8871874809265137, "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": 979.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 466.0625, "completions/mean_terminated_length": 466.0625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 1.3168, "grad_norm": 0.052106019109487534, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32682267.0, "reward": 1.0349509716033936, "reward_std": 0.39237916469573975, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5698999762535095, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9378124475479126, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 589.828125, "completions/mean_terminated_length": 582.9365234375, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 1.3176, "grad_norm": 0.04524199664592743, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32736976.0, "reward": 0.9792704582214355, "reward_std": 0.4265438914299011, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5991640090942383, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.780750036239624, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 459.734375, "completions/mean_terminated_length": 367.3999938964844, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.3184, "grad_norm": 0.06355322897434235, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32782191.0, "reward": 0.9855852723121643, "reward_std": 0.29256361722946167, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6117937564849854, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7559374570846558, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 466.859375, "completions/mean_terminated_length": 458.0158996582031, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.3192, "grad_norm": 0.0719570443034172, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32828054.0, "reward": 1.0045862197875977, "reward_std": 0.4392656683921814, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5716702938079834, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8884375095367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 634.25, "completions/mean_terminated_length": 608.2667236328125, "completions/min_length": 306.0, "completions/min_terminated_length": 306.0, "epoch": 1.32, "grad_norm": 0.03859522566199303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32885958.0, "reward": 0.7902610898017883, "reward_std": 0.3604658544063568, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.36177030205726624, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8829687833786011, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 760.0, "completions/mean_length": 482.875, "completions/mean_terminated_length": 474.2857360839844, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.3208, "grad_norm": 0.06287592649459839, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 32933918.0, "reward": 0.9741719961166382, "reward_std": 0.4358474612236023, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5889672040939331, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8254687786102295, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 478.734375, "completions/mean_terminated_length": 461.1451416015625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.3216, "grad_norm": 0.036894649267196655, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 32980253.0, "reward": 0.8229930400848389, "reward_std": 0.28990721702575684, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.45848435163497925, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7512500286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 597.046875, "completions/mean_terminated_length": 590.2698974609375, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 1.3224, "grad_norm": 0.046219952404499054, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33033840.0, "reward": 0.9202321767807007, "reward_std": 0.44535350799560547, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.49671247601509094, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8739062547683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 464.890625, "completions/mean_terminated_length": 456.0158996582031, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.3232, "grad_norm": 0.03772368282079697, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33079273.0, "reward": 0.9037603139877319, "reward_std": 0.3736369013786316, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.49501875042915344, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8637499809265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 980.0, "completions/mean_length": 446.125, "completions/mean_terminated_length": 436.952392578125, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 1.324, "grad_norm": 0.06674040108919144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33123057.0, "reward": 0.9405087232589722, "reward_std": 0.4270162582397461, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5216406583786011, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.862500011920929, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 479.421875, "completions/mean_terminated_length": 433.27117919921875, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 1.3248, "grad_norm": 0.05303975194692612, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33170204.0, "reward": 1.0232915878295898, "reward_std": 0.4459725022315979, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5778312683105469, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8840625286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 547.328125, "completions/mean_terminated_length": 539.761962890625, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 1.3256000000000001, "grad_norm": 0.03954806551337242, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33223265.0, "reward": 0.8232237696647644, "reward_std": 0.34794431924819946, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3651953339576721, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9445313215255737, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 434.46875, "completions/mean_terminated_length": 425.11114501953125, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.3264, "grad_norm": 0.05418756976723671, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33267247.0, "reward": 0.9753197431564331, "reward_std": 0.3443257212638855, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5756374597549438, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8459374904632568, "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": 799.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 349.625, "completions/mean_terminated_length": 349.625, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 1.3272, "grad_norm": 0.053901251405477524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33304423.0, "reward": 1.1362042427062988, "reward_std": 0.2788000702857971, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6630312204360962, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9406249523162842, "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": 793.0, "completions/max_terminated_length": 793.0, "completions/mean_length": 371.6875, "completions/mean_terminated_length": 371.6875, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.328, "grad_norm": 0.06446096301078796, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33345059.0, "reward": 0.7911682724952698, "reward_std": 0.3200055956840515, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.37920936942100525, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8415625095367432, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 545.828125, "completions/mean_terminated_length": 522.3114624023438, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.3288, "grad_norm": 0.052012670785188675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33395400.0, "reward": 0.8222070336341858, "reward_std": 0.39300259947776794, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.47253745794296265, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7357812523841858, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 579.140625, "completions/mean_terminated_length": 524.5087890625, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 1.3296000000000001, "grad_norm": 0.04575250297784805, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33450513.0, "reward": 0.8460532426834106, "reward_std": 0.3644160032272339, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.426479697227478, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8420312404632568, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 680.34375, "completions/mean_terminated_length": 631.25, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 1.3304, "grad_norm": 0.02601347304880619, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33512567.0, "reward": 0.6503024697303772, "reward_std": 0.3411686420440674, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.33185312151908875, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6753125190734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 540.296875, "completions/mean_terminated_length": 524.6935424804688, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 1.3312, "grad_norm": 0.040369629859924316, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33565290.0, "reward": 0.7732226848602295, "reward_std": 0.2626665234565735, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.46831873059272766, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.6756249666213989, "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": 743.0, "completions/max_terminated_length": 743.0, "completions/mean_length": 528.984375, "completions/mean_terminated_length": 528.984375, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.332, "grad_norm": 0.024719716981053352, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33615273.0, "reward": 0.8765665888786316, "reward_std": 0.20643103122711182, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.47188127040863037, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8600000143051147, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 463.296875, "completions/mean_terminated_length": 435.7212829589844, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.3328, "grad_norm": 0.1038026437163353, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33659868.0, "reward": 0.9780407547950745, "reward_std": 0.4652441740036011, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5342047214508057, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.87890625, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 518.265625, "completions/mean_terminated_length": 493.3934020996094, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.3336000000000001, "grad_norm": 0.044911306351423264, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33708701.0, "reward": 1.2027978897094727, "reward_std": 0.38924527168273926, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.827468752861023, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7625000476837158, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 474.171875, "completions/mean_terminated_length": 465.4444580078125, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.3344, "grad_norm": 0.03973904997110367, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33754488.0, "reward": 1.0364291667938232, "reward_std": 0.3427441418170929, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5884812474250793, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9071875214576721, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 511.28125, "completions/mean_terminated_length": 486.0655517578125, "completions/min_length": 301.0, "completions/min_terminated_length": 301.0, "epoch": 1.3352, "grad_norm": 0.039354510605335236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33803514.0, "reward": 1.229191541671753, "reward_std": 0.3090945780277252, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7708812355995178, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8856250047683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 891.0, "completions/mean_length": 527.953125, "completions/mean_terminated_length": 511.95159912109375, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 1.336, "grad_norm": 0.0562637560069561, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33854231.0, "reward": 0.8232258558273315, "reward_std": 0.42145341634750366, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4433249831199646, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7654687166213989, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 518.34375, "completions/mean_terminated_length": 466.03448486328125, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 1.3368, "grad_norm": 0.05009065195918083, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 33903437.0, "reward": 1.1331274509429932, "reward_std": 0.39378032088279724, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.7193781137466431, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8140625357627869, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 506.9375, "completions/mean_terminated_length": 463.11865234375, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.3376000000000001, "grad_norm": 0.07575641572475433, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 33951785.0, "reward": 0.9618704319000244, "reward_std": 0.4807930290699005, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5799890756607056, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7757812738418579, "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": 0.0, "completions/max_length": 850.0, "completions/max_terminated_length": 850.0, "completions/mean_length": 513.8125, "completions/mean_terminated_length": 513.8125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.3384, "grad_norm": 0.053672466427087784, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34000397.0, "reward": 0.7773610949516296, "reward_std": 0.2882513999938965, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.33597031235694885, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9207812547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 522.328125, "completions/mean_terminated_length": 506.1451416015625, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 1.3392, "grad_norm": 0.06478086858987808, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34049698.0, "reward": 0.9551875591278076, "reward_std": 0.41302984952926636, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.5978734493255615, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7598437070846558, "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": 1000.0, "completions/mean_length": 503.796875, "completions/mean_terminated_length": 469.11669921875, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 1.34, "grad_norm": 0.045184001326560974, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34099525.0, "reward": 0.8279790878295898, "reward_std": 0.377496600151062, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4215812385082245, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8431249856948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 673.0, "completions/mean_length": 409.03125, "completions/mean_terminated_length": 399.2698669433594, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.3408, "grad_norm": 0.05875379964709282, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34140535.0, "reward": 0.993781328201294, "reward_std": 0.41719216108322144, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.6125609874725342, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8051562905311584, "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": 962.0, "completions/mean_length": 588.75, "completions/mean_terminated_length": 559.7333374023438, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 1.3416000000000001, "grad_norm": 0.0341002494096756, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34195335.0, "reward": 0.7163448333740234, "reward_std": 0.31418460607528687, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3233124911785126, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8206250071525574, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 438.0, "completions/mean_terminated_length": 419.0967712402344, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.3424, "grad_norm": 0.0676739439368248, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34238743.0, "reward": 1.0506691932678223, "reward_std": 0.4817992150783539, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.585710883140564, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9173437356948853, "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": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 407.015625, "completions/mean_terminated_length": 407.015625, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 1.3432, "grad_norm": 0.05877132713794708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34282088.0, "reward": 1.0086729526519775, "reward_std": 0.42820870876312256, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5329687595367432, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9546875357627869, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 437.78125, "completions/mean_terminated_length": 408.9508056640625, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 1.3439999999999999, "grad_norm": 0.06427907198667526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34327546.0, "reward": 1.0461041927337646, "reward_std": 0.4204477369785309, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5922062397003174, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9043750166893005, "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": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 484.515625, "completions/mean_terminated_length": 484.515625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.3448, "grad_norm": 0.05630764737725258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34376635.0, "reward": 0.7754197120666504, "reward_std": 0.2662498652935028, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.3320874869823456, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9143750071525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 417.4375, "completions/mean_terminated_length": 407.8095397949219, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.3456000000000001, "grad_norm": 0.07347879558801651, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34418871.0, "reward": 1.0054705142974854, "reward_std": 0.41730448603630066, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5734390616416931, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8810937404632568, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 573.8125, "completions/mean_terminated_length": 543.800048828125, "completions/min_length": 284.0, "completions/min_terminated_length": 284.0, "epoch": 1.3464, "grad_norm": 0.03315845504403114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34472619.0, "reward": 0.6614034175872803, "reward_std": 0.3220205307006836, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.24467968940734863, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8589062690734863, "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": 900.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 365.9375, "completions/mean_terminated_length": 365.9375, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "epoch": 1.3472, "grad_norm": 0.08363376557826996, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34512231.0, "reward": 1.025038480758667, "reward_std": 0.44994109869003296, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5500749349594116, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9533734321594238, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 496.953125, "completions/mean_terminated_length": 479.95159912109375, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.3479999999999999, "grad_norm": 0.06184015050530434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34559652.0, "reward": 1.0669867992401123, "reward_std": 0.3869190812110901, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6495969295501709, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8537499904632568, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 622.0, "completions/mean_length": 395.359375, "completions/mean_terminated_length": 375.08062744140625, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.3488, "grad_norm": 0.07741474360227585, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34600475.0, "reward": 1.0007878541946411, "reward_std": 0.43930625915527344, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5484484434127808, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9039062261581421, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 500.734375, "completions/mean_terminated_length": 474.9999694824219, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 1.3496000000000001, "grad_norm": 0.049500156193971634, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34648954.0, "reward": 0.85651034116745, "reward_std": 0.4576595723628998, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4317687153816223, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8517187833786011, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 595.875, "completions/mean_terminated_length": 567.3333740234375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 1.3504, "grad_norm": 0.0507885217666626, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34702610.0, "reward": 0.7770369052886963, "reward_std": 0.35096418857574463, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3665718734264374, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8503124713897705, "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": 0.0, "completions/max_length": 985.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 462.0, "completions/mean_terminated_length": 462.0, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 1.3512, "grad_norm": 0.05478418990969658, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34748306.0, "reward": 0.8220017552375793, "reward_std": 0.395066499710083, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3783765733242035, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9242187738418579, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 612.15625, "completions/mean_terminated_length": 569.5516967773438, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.3519999999999999, "grad_norm": 0.03469787910580635, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 34805372.0, "reward": 0.8193328380584717, "reward_std": 0.26287633180618286, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4980390667915344, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7214062213897705, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 497.109375, "completions/mean_terminated_length": 432.40350341796875, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.3528, "grad_norm": 0.08165766298770905, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34854611.0, "reward": 0.8404110670089722, "reward_std": 0.4422893822193146, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4620703160762787, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7456250190734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 492.4375, "completions/mean_terminated_length": 484.0000305175781, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 1.3536000000000001, "grad_norm": 0.06091555207967758, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34902239.0, "reward": 0.9718768000602722, "reward_std": 0.36679357290267944, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5062515735626221, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9326562285423279, "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": 818.0, "completions/max_terminated_length": 818.0, "completions/mean_length": 379.515625, "completions/mean_terminated_length": 379.515625, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 1.3544, "grad_norm": 0.07077722996473312, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34943664.0, "reward": 0.9789690971374512, "reward_std": 0.39371347427368164, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5360609292984009, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.89453125, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 569.59375, "completions/mean_terminated_length": 495.2363586425781, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.3552, "grad_norm": 0.05527452006936073, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 34996182.0, "reward": 0.7335524559020996, "reward_std": 0.3088521957397461, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.40460312366485596, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.6884374618530273, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 765.0, "completions/mean_length": 447.5625, "completions/mean_terminated_length": 438.4127197265625, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.3559999999999999, "grad_norm": 0.051831405609846115, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35040394.0, "reward": 1.0745024681091309, "reward_std": 0.39635926485061646, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6646281480789185, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8259375095367432, "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": 0.0, "completions/max_length": 717.0, "completions/max_terminated_length": 717.0, "completions/mean_length": 398.21875, "completions/mean_terminated_length": 398.21875, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 1.3568, "grad_norm": 0.09023384749889374, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35082936.0, "reward": 1.1505417823791504, "reward_std": 0.4223599433898926, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6760812401771545, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9495312571525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 414.484375, "completions/mean_terminated_length": 404.8095397949219, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.3576, "grad_norm": 0.07654442638158798, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35125735.0, "reward": 1.0269954204559326, "reward_std": 0.3133258819580078, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.553989052772522, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9429687857627869, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 618.359375, "completions/mean_terminated_length": 583.9830322265625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.3584, "grad_norm": 0.04902392998337746, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35181790.0, "reward": 0.9798542261123657, "reward_std": 0.46264195442199707, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.6003312468528748, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7915624976158142, "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": 982.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 479.125, "completions/mean_terminated_length": 479.125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.3592, "grad_norm": 0.050593916326761246, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35230198.0, "reward": 0.9785383939743042, "reward_std": 0.35751864314079285, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5820749998092651, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8231250047683716, "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": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 455.90625, "completions/mean_terminated_length": 455.90625, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 1.3599999999999999, "grad_norm": 0.10827641189098358, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35276176.0, "reward": 0.881873607635498, "reward_std": 0.34532374143600464, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4824953079223633, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.83984375, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 399.765625, "completions/mean_terminated_length": 389.857177734375, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 1.3608, "grad_norm": 0.06377716362476349, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35317729.0, "reward": 0.9493885636329651, "reward_std": 0.3742981553077698, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47690001130104065, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9471874833106995, "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": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 363.859375, "completions/mean_terminated_length": 363.859375, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 1.3616, "grad_norm": 0.051514893770217896, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35356776.0, "reward": 1.100878357887268, "reward_std": 0.28243234753608704, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6236296892166138, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9507812261581421, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 566.234375, "completions/mean_terminated_length": 527.440673828125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.3624, "grad_norm": 0.02587469480931759, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35410775.0, "reward": 0.9094827175140381, "reward_std": 0.22866995632648468, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.6158390045166016, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.6223437190055847, "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": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 387.578125, "completions/mean_terminated_length": 387.578125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.3632, "grad_norm": 0.07992936670780182, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35453580.0, "reward": 1.0490531921386719, "reward_std": 0.38452035188674927, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.61372971534729, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8923437595367432, "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": 652.0, "completions/max_terminated_length": 652.0, "completions/mean_length": 369.0625, "completions/mean_terminated_length": 369.0625, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.3639999999999999, "grad_norm": 0.05996957793831825, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35496144.0, "reward": 0.8866518139839172, "reward_std": 0.38429245352745056, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4139265716075897, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.95703125, "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": 0.0, "completions/max_length": 905.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 413.640625, "completions/mean_terminated_length": 413.640625, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.3648, "grad_norm": 0.05184457078576088, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35538201.0, "reward": 0.7757753133773804, "reward_std": 0.29059791564941406, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.37967342138290405, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8285937309265137, "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": 0.0, "completions/max_length": 898.0, "completions/max_terminated_length": 898.0, "completions/mean_length": 410.421875, "completions/mean_terminated_length": 410.421875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.3656, "grad_norm": 0.07535193860530853, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35579684.0, "reward": 1.2069730758666992, "reward_std": 0.23062875866889954, "rewards/accuracy_reward": 0.6875, "rewards/brier_reward": 0.7420687675476074, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9370312690734863, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 411.75, "completions/mean_terminated_length": 381.6393127441406, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.3664, "grad_norm": 0.07159451395273209, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35622740.0, "reward": 0.977420449256897, "reward_std": 0.3648113012313843, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5329641103744507, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8967187404632568, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 533.421875, "completions/mean_terminated_length": 500.7167053222656, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.3672, "grad_norm": 0.04395494982600212, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35673119.0, "reward": 0.8942517638206482, "reward_std": 0.35038071870803833, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5072515606880188, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7960937023162842, "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": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 378.5625, "completions/mean_terminated_length": 378.5625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.3679999999999999, "grad_norm": 0.05472799763083458, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35714099.0, "reward": 0.9967352747917175, "reward_std": 0.33466461300849915, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.6497187614440918, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7359374761581421, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 531.890625, "completions/mean_terminated_length": 524.0794067382812, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 1.3688, "grad_norm": 0.1382414698600769, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35766844.0, "reward": 1.0045790672302246, "reward_std": 0.45021653175354004, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5716562271118164, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9009374976158142, "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": 0.0, "completions/max_length": 623.0, "completions/max_terminated_length": 623.0, "completions/mean_length": 221.125, "completions/mean_terminated_length": 221.125, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 1.3696, "grad_norm": 0.09852050244808197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35796004.0, "reward": 1.2255465984344482, "reward_std": 0.27596408128738403, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7323405742645264, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9749844074249268, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 584.5, "completions/mean_terminated_length": 530.5263061523438, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.3704, "grad_norm": 0.06163449212908745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35849700.0, "reward": 0.69273442029953, "reward_std": 0.39886438846588135, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.38546717166900635, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.681093692779541, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 349.265625, "completions/mean_terminated_length": 338.5555725097656, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "epoch": 1.3712, "grad_norm": 0.11239340901374817, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35889317.0, "reward": 0.9564628005027771, "reward_std": 0.4066208600997925, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4910484552383423, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9364062547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 557.484375, "completions/mean_terminated_length": 542.4354858398438, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 1.3719999999999999, "grad_norm": 0.03609195724129677, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 35940548.0, "reward": 0.7208197116851807, "reward_std": 0.3188169598579407, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3166375160217285, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8645312190055847, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 455.984375, "completions/mean_terminated_length": 446.9682922363281, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.3728, "grad_norm": 0.046290092170238495, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 35984995.0, "reward": 1.0463485717773438, "reward_std": 0.29425370693206787, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6395702958106995, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.844531238079071, "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": 0.0, "completions/max_length": 798.0, "completions/max_terminated_length": 798.0, "completions/mean_length": 479.265625, "completions/mean_terminated_length": 479.265625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.3736, "grad_norm": 0.04543505981564522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36032436.0, "reward": 0.9010353088378906, "reward_std": 0.36689943075180054, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.44269371032714844, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9221875667572021, "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": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 441.1875, "completions/mean_terminated_length": 441.1875, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 1.3744, "grad_norm": 0.08419936150312424, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36076960.0, "reward": 1.1203736066818237, "reward_std": 0.41512560844421387, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6938703060150146, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.87109375, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 722.0, "completions/mean_length": 449.65625, "completions/mean_terminated_length": 440.5397033691406, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.3752, "grad_norm": 0.06909216195344925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36122234.0, "reward": 0.8336033225059509, "reward_std": 0.3611493706703186, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.38595467805862427, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9167187213897705, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 547.625, "completions/mean_terminated_length": 524.1967163085938, "completions/min_length": 262.0, "completions/min_terminated_length": 262.0, "epoch": 1.376, "grad_norm": 0.05562763288617134, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36174786.0, "reward": 0.9022611379623413, "reward_std": 0.4785464406013489, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.49202027916908264, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.827343761920929, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 499.90625, "completions/mean_terminated_length": 464.9667053222656, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 1.3768, "grad_norm": 0.032576508820056915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36223132.0, "reward": 1.1719008684158325, "reward_std": 0.2536951005458832, "rewards/accuracy_reward": 0.671875, "rewards/brier_reward": 0.7344250082969666, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8578124642372131, "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": 920.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 465.015625, "completions/mean_terminated_length": 465.015625, "completions/min_length": 132.0, "completions/min_terminated_length": 132.0, "epoch": 1.3776, "grad_norm": 0.03294955939054489, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36269709.0, "reward": 0.8892090916633606, "reward_std": 0.25123149156570435, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.41904062032699585, "rewards/confidence_one_or_zero": 0.375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9590625166893005, "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": 711.0, "completions/max_terminated_length": 711.0, "completions/mean_length": 399.390625, "completions/mean_terminated_length": 399.390625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 1.3784, "grad_norm": 0.0677311047911644, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36310646.0, "reward": 0.8716479539871216, "reward_std": 0.37966954708099365, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.41516876220703125, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9334374666213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 530.390625, "completions/mean_terminated_length": 514.4677124023438, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 1.3792, "grad_norm": 0.04655800387263298, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36360527.0, "reward": 0.6469854116439819, "reward_std": 0.3273572325706482, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.21584375202655792, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8979687690734863, "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": 858.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 365.609375, "completions/mean_terminated_length": 365.609375, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.38, "grad_norm": 0.02537403628230095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36399814.0, "reward": 0.8115581274032593, "reward_std": 0.13636498153209686, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.35748907923698425, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9381250143051147, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 548.984375, "completions/mean_terminated_length": 517.3167114257812, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.3808, "grad_norm": 0.04474574327468872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36451125.0, "reward": 0.7644057869911194, "reward_std": 0.37509337067604065, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3256843686103821, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8940625190734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 508.640625, "completions/mean_terminated_length": 492.01611328125, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 1.3816, "grad_norm": 0.09323311597108841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36499550.0, "reward": 1.0503562688827515, "reward_std": 0.4248843789100647, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.663210928440094, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7907812595367432, "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": 788.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 475.953125, "completions/mean_terminated_length": 475.953125, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 1.3824, "grad_norm": 0.06918696314096451, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36546091.0, "reward": 1.051580548286438, "reward_std": 0.3977513611316681, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6031594276428223, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8862500190734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 821.0, "completions/mean_length": 491.125, "completions/mean_terminated_length": 482.66668701171875, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.3832, "grad_norm": 0.04617975279688835, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36593955.0, "reward": 1.0056719779968262, "reward_std": 0.324300080537796, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.6050921678543091, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8270312547683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 775.0, "completions/mean_length": 430.921875, "completions/mean_terminated_length": 421.5079650878906, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 1.384, "grad_norm": 0.05836394801735878, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36637502.0, "reward": 0.8375376462936401, "reward_std": 0.37802696228027344, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.37819841504096985, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9373437166213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 834.0, "completions/mean_length": 488.234375, "completions/mean_terminated_length": 470.95159912109375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 1.3848, "grad_norm": 0.07062748819589615, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36684397.0, "reward": 1.049519658088684, "reward_std": 0.4563475549221039, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5834125280380249, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9034374952316284, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 920.0, "completions/mean_length": 345.34375, "completions/mean_terminated_length": 334.5714416503906, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "epoch": 1.3856, "grad_norm": 0.07773531973361969, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36723075.0, "reward": 0.8799854516983032, "reward_std": 0.42431995272636414, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.43184375762939453, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9240624904632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 556.078125, "completions/mean_terminated_length": 548.6508178710938, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.3864, "grad_norm": 0.035489995032548904, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36777080.0, "reward": 0.8393046855926514, "reward_std": 0.3265388011932373, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.5223578214645386, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.6901562213897705, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 618.1875, "completions/mean_terminated_length": 551.7817993164062, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 1.3872, "grad_norm": 0.04265666753053665, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36834052.0, "reward": 0.7626780867576599, "reward_std": 0.3759097456932068, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.38472965359687805, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7760937213897705, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 985.0, "completions/mean_length": 475.828125, "completions/mean_terminated_length": 429.3728942871094, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 1.388, "grad_norm": 0.04816456511616707, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 36880601.0, "reward": 0.9834628105163574, "reward_std": 0.28127121925354004, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5762984752655029, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.83203125, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 442.125, "completions/mean_terminated_length": 413.5081787109375, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.3888, "grad_norm": 0.07075648754835129, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36925153.0, "reward": 1.1388227939605713, "reward_std": 0.4102310538291931, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6995187401771545, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8709374666213989, "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": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 370.375, "completions/mean_terminated_length": 370.375, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 1.3896, "grad_norm": 0.05274735391139984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 36964457.0, "reward": 1.3472760915756226, "reward_std": 0.26736679673194885, "rewards/accuracy_reward": 0.84375, "rewards/brier_reward": 0.8508000373840332, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9650000333786011, "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": 648.0, "completions/max_terminated_length": 648.0, "completions/mean_length": 357.25, "completions/mean_terminated_length": 357.25, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.3904, "grad_norm": 0.09067439287900925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37003225.0, "reward": 1.1201674938201904, "reward_std": 0.4022870659828186, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.646582841873169, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9479687213897705, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 832.0, "completions/mean_length": 442.953125, "completions/mean_terminated_length": 404.2166748046875, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 1.3912, "grad_norm": 0.06965970247983932, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37046630.0, "reward": 0.7444728016853333, "reward_std": 0.4475051760673523, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.28581875562667847, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8596874475479126, "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": 0.0, "completions/max_length": 939.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 524.515625, "completions/mean_terminated_length": 524.515625, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.392, "grad_norm": 0.030781330540776253, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37096167.0, "reward": 0.8360104560852051, "reward_std": 0.22489622235298157, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.42201876640319824, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8574999570846558, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 482.546875, "completions/mean_terminated_length": 465.08062744140625, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 1.3928, "grad_norm": 0.09966684877872467, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37142154.0, "reward": 0.9423503279685974, "reward_std": 0.4709184467792511, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47844842076301575, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9304687976837158, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 548.359375, "completions/mean_terminated_length": 499.1551818847656, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 1.3936, "grad_norm": 0.051400456577539444, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37193713.0, "reward": 0.7181617021560669, "reward_std": 0.35323333740234375, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.38944685459136963, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7184374928474426, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 373.203125, "completions/mean_terminated_length": 362.873046875, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.3944, "grad_norm": 0.0694802924990654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37232910.0, "reward": 0.8420979976654053, "reward_std": 0.28386449813842773, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.43419376015663147, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8481249809265137, "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": 969.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 372.640625, "completions/mean_terminated_length": 372.640625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.3952, "grad_norm": 0.055784136056900024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37273047.0, "reward": 1.1106908321380615, "reward_std": 0.3780170679092407, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6588796973228455, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9039062857627869, "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": 1016.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 522.0, "completions/mean_terminated_length": 522.0, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 1.396, "grad_norm": 0.05733221396803856, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37322439.0, "reward": 0.798697829246521, "reward_std": 0.31377917528152466, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3630187511444092, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9015624523162842, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 493.21875, "completions/mean_terminated_length": 476.0967712402344, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 1.3968, "grad_norm": 0.05963108316063881, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37369397.0, "reward": 0.7366968393325806, "reward_std": 0.3198680877685547, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.48901715874671936, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.5692187547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 754.0, "completions/mean_length": 424.21875, "completions/mean_terminated_length": 404.8709411621094, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.3976, "grad_norm": 0.042311299592256546, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37412275.0, "reward": 0.8826384544372559, "reward_std": 0.3005449175834656, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4215250015258789, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9287500381469727, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 574.5, "completions/mean_terminated_length": 519.2982788085938, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.3984, "grad_norm": 0.0628160610795021, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37465059.0, "reward": 0.694198489189148, "reward_std": 0.3901869058609009, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.3571453094482422, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7176562547683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 750.0, "completions/mean_length": 456.328125, "completions/mean_terminated_length": 438.01611328125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 1.3992, "grad_norm": 0.0680345743894577, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37512488.0, "reward": 0.7802063226699829, "reward_std": 0.3882943391799927, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.38853591680526733, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8170312643051147, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 475.390625, "completions/mean_terminated_length": 457.69354248046875, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 1.4, "grad_norm": 0.03058202564716339, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37559777.0, "reward": 0.5641963481903076, "reward_std": 0.15242472290992737, "rewards/accuracy_reward": 0.046875, "rewards/brier_reward": 0.1283906251192093, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9096874594688416, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 595.796875, "completions/mean_terminated_length": 589.0000610351562, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 1.4008, "grad_norm": 0.044084466993808746, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37615732.0, "reward": 0.7755993604660034, "reward_std": 0.310820609331131, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.4730718731880188, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.6664062738418579, "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": 0.0, "completions/max_length": 924.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 474.5625, "completions/mean_terminated_length": 474.5625, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.4016, "grad_norm": 0.0645517036318779, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37662488.0, "reward": 1.0566120147705078, "reward_std": 0.3833525478839874, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6132218837738037, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9228124618530273, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 498.9375, "completions/mean_terminated_length": 473.1147155761719, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.4024, "grad_norm": 0.06690462678670883, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37710164.0, "reward": 1.115911841392517, "reward_std": 0.42260313034057617, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6536968946456909, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9090625047683716, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 620.71875, "completions/mean_terminated_length": 571.1929931640625, "completions/min_length": 325.0, "completions/min_terminated_length": 325.0, "epoch": 1.4032, "grad_norm": 0.026705302298069, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37766258.0, "reward": 0.8586828708648682, "reward_std": 0.3057166635990143, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4673640727996826, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7951562404632568, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 526.59375, "completions/mean_terminated_length": 502.1311340332031, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.404, "grad_norm": 0.0922914445400238, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37815784.0, "reward": 1.0185571908950806, "reward_std": 0.3917151391506195, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5839874744415283, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8864062428474426, "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": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 462.625, "completions/mean_terminated_length": 462.625, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 1.4048, "grad_norm": 0.05745416134595871, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37862672.0, "reward": 0.981083869934082, "reward_std": 0.3677506446838379, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5402906537055969, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9221875071525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 791.0, "completions/mean_length": 435.515625, "completions/mean_terminated_length": 426.17462158203125, "completions/min_length": 282.0, "completions/min_terminated_length": 282.0, "epoch": 1.4056, "grad_norm": 0.05167049914598465, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 37910177.0, "reward": 0.962786078453064, "reward_std": 0.3453923463821411, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5505703091621399, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8435937166213989, "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": 783.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 387.046875, "completions/mean_terminated_length": 387.046875, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.4064, "grad_norm": 0.06648380309343338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 37951428.0, "reward": 0.7258669137954712, "reward_std": 0.18913526833057404, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26423123478889465, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9490625262260437, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 673.765625, "completions/mean_terminated_length": 623.732177734375, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 1.4072, "grad_norm": 0.04106992483139038, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38013621.0, "reward": 0.6582031846046448, "reward_std": 0.32176727056503296, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26952967047691345, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8139062523841858, "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": 802.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 379.828125, "completions/mean_terminated_length": 379.828125, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.408, "grad_norm": 0.058962415903806686, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38053962.0, "reward": 1.1579556465148926, "reward_std": 0.36065882444381714, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6909093856811523, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9262499809265137, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 651.71875, "completions/mean_terminated_length": 582.7777709960938, "completions/min_length": 310.0, "completions/min_terminated_length": 310.0, "epoch": 1.4088, "grad_norm": 0.036083515733480453, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38112056.0, "reward": 0.7632491588592529, "reward_std": 0.3711729049682617, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.46399685740470886, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.6564062833786011, "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": 0.0, "completions/max_length": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 430.75, "completions/mean_terminated_length": 430.75, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 1.4096, "grad_norm": 0.09401793032884598, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38156744.0, "reward": 1.0756330490112305, "reward_std": 0.34187865257263184, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6200140714645386, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.921875, "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": 978.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 556.65625, "completions/mean_terminated_length": 556.65625, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 1.4104, "grad_norm": 0.022671593353152275, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38209298.0, "reward": 0.8289720416069031, "reward_std": 0.19630518555641174, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.43919217586517334, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8467187285423279, "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": 991.0, "completions/mean_length": 494.234375, "completions/mean_terminated_length": 458.91668701171875, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.4112, "grad_norm": 0.057079214602708817, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38256945.0, "reward": 1.07170569896698, "reward_std": 0.39247363805770874, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6590343713760376, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8356249928474426, "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": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 476.625, "completions/mean_terminated_length": 476.625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 1.412, "grad_norm": 0.016786746680736542, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38304873.0, "reward": 0.8108729124069214, "reward_std": 0.10272090137004852, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.35611873865127563, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9359374642372131, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 449.875, "completions/mean_terminated_length": 421.6393127441406, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.4128, "grad_norm": 0.08579767495393753, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38351057.0, "reward": 0.9571767449378967, "reward_std": 0.41595298051834106, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5081015825271606, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9032812118530273, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 501.515625, "completions/mean_terminated_length": 484.6612854003906, "completions/min_length": 214.0, "completions/min_terminated_length": 214.0, "epoch": 1.4136, "grad_norm": 0.06476894021034241, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38399250.0, "reward": 0.9346072673797607, "reward_std": 0.38747596740722656, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4942125082015991, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8918750286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 741.0, "completions/mean_length": 342.828125, "completions/mean_terminated_length": 332.0158996582031, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.4144, "grad_norm": 0.08053100854158401, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38437655.0, "reward": 0.831126868724823, "reward_std": 0.31531935930252075, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3653765320777893, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9442187547683716, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 557.4375, "completions/mean_terminated_length": 526.3333740234375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.4152, "grad_norm": 0.05532347783446312, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38490355.0, "reward": 0.968043327331543, "reward_std": 0.3906235694885254, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5142093896865845, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 641.4375, "completions/mean_terminated_length": 629.0967407226562, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 1.416, "grad_norm": 0.035769060254096985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38548623.0, "reward": 0.693203330039978, "reward_std": 0.3524077832698822, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.29265469312667847, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8709374666213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 554.703125, "completions/mean_terminated_length": 539.5645141601562, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.4168, "grad_norm": 0.028478246182203293, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38599836.0, "reward": 0.8120306730270386, "reward_std": 0.2501620650291443, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.45218437910079956, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7684375047683716, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 969.0, "completions/mean_length": 539.171875, "completions/mean_terminated_length": 515.3278198242188, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 1.4176, "grad_norm": 0.05599226430058479, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38651399.0, "reward": 1.1591923236846924, "reward_std": 0.44399428367614746, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7402577996253967, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8604687452316284, "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": 0.0, "completions/max_length": 989.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 485.265625, "completions/mean_terminated_length": 485.265625, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 1.4184, "grad_norm": 0.06413860619068146, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38697688.0, "reward": 1.0789151191711426, "reward_std": 0.352525532245636, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6109530925750732, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.934999942779541, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 607.671875, "completions/mean_terminated_length": 548.1964721679688, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.4192, "grad_norm": 0.07347487658262253, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38754195.0, "reward": 0.6690922975540161, "reward_std": 0.40039294958114624, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.32255780696868896, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7387499809265137, "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": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 423.046875, "completions/mean_terminated_length": 423.046875, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 1.42, "grad_norm": 0.028948578983545303, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38796070.0, "reward": 1.3023300170898438, "reward_std": 0.19908565282821655, "rewards/accuracy_reward": 0.78125, "rewards/brier_reward": 0.823407769203186, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.944515585899353, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 478.03125, "completions/mean_terminated_length": 451.1802978515625, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 1.4208, "grad_norm": 0.09446964412927628, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 38843384.0, "reward": 0.9849267601966858, "reward_std": 0.42170560359954834, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.532351553440094, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9054687023162842, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 493.53125, "completions/mean_terminated_length": 467.4425964355469, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.4216, "grad_norm": 0.0876612663269043, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38892554.0, "reward": 1.0344853401184082, "reward_std": 0.5220067501068115, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.584593653678894, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8877812623977661, "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": 0.0, "completions/max_length": 970.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 518.71875, "completions/mean_terminated_length": 518.71875, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 1.4224, "grad_norm": 0.06663181632757187, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38944152.0, "reward": 0.9089557528495789, "reward_std": 0.34944966435432434, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4429093599319458, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9535937309265137, "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": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 546.796875, "completions/mean_terminated_length": 546.796875, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 1.4232, "grad_norm": 0.04747254028916359, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 38996059.0, "reward": 1.0077173709869385, "reward_std": 0.3814852237701416, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5623078346252441, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8964062929153442, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 724.0, "completions/mean_length": 509.78125, "completions/mean_terminated_length": 484.4917907714844, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 1.424, "grad_norm": 0.04587574303150177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39044333.0, "reward": 0.7894735336303711, "reward_std": 0.3206072449684143, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3758203089237213, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8695312142372131, "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": 0.0, "completions/max_length": 772.0, "completions/max_terminated_length": 772.0, "completions/mean_length": 475.5, "completions/mean_terminated_length": 475.5, "completions/min_length": 337.0, "completions/min_terminated_length": 337.0, "epoch": 1.4248, "grad_norm": 0.06397241353988647, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39092109.0, "reward": 0.9907649755477905, "reward_std": 0.4295171797275543, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5596531629562378, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9043749570846558, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 436.359375, "completions/mean_terminated_length": 407.458984375, "completions/min_length": 243.0, "completions/min_terminated_length": 243.0, "epoch": 1.4256, "grad_norm": 0.04567555710673332, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39136756.0, "reward": 1.0365290641784668, "reward_std": 0.33330219984054565, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5886812210083008, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8928124904632568, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 556.484375, "completions/mean_terminated_length": 533.4917602539062, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 1.4264000000000001, "grad_norm": 0.04744448512792587, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39187811.0, "reward": 0.8573577404022217, "reward_std": 0.40354570746421814, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.5740890502929688, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.6314061880111694, "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": 907.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 438.15625, "completions/mean_terminated_length": 438.15625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.4272, "grad_norm": 0.03913486748933792, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39231501.0, "reward": 0.9215233325958252, "reward_std": 0.25192558765411377, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.6399202942848206, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6560937166213989, "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": 828.0, "completions/max_terminated_length": 828.0, "completions/mean_length": 423.171875, "completions/mean_terminated_length": 423.171875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.428, "grad_norm": 0.062362223863601685, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39276712.0, "reward": 0.924757182598114, "reward_std": 0.2604280114173889, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.52138751745224, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8487499952316284, "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": 0.0, "completions/max_length": 1007.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 499.0, "completions/mean_terminated_length": 499.0, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.4288, "grad_norm": 0.06461799144744873, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39326376.0, "reward": 0.8373681902885437, "reward_std": 0.3657475709915161, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.42473435401916504, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8856250047683716, "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": 825.0, "completions/max_terminated_length": 825.0, "completions/mean_length": 436.28125, "completions/mean_terminated_length": 436.28125, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.4296, "grad_norm": 0.07901297509670258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39370810.0, "reward": 0.8850136995315552, "reward_std": 0.42932790517807007, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.41065001487731934, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9571875333786011, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 744.0, "completions/mean_length": 430.09375, "completions/mean_terminated_length": 420.66668701171875, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.4304000000000001, "grad_norm": 0.05678265169262886, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39416944.0, "reward": 0.9991480112075806, "reward_std": 0.43536216020584106, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5295437574386597, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9479687213897705, "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": 0.0, "completions/max_length": 963.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 579.0, "completions/mean_terminated_length": 579.0, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 1.4312, "grad_norm": 0.04153214767575264, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39473984.0, "reward": 0.8446071743965149, "reward_std": 0.272733598947525, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39233750104904175, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9459375143051147, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 632.5625, "completions/mean_terminated_length": 592.0689697265625, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.432, "grad_norm": 0.05153611674904823, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39532132.0, "reward": 0.8904563188552856, "reward_std": 0.34539932012557983, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.5152859687805176, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7751562595367432, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 440.15625, "completions/mean_terminated_length": 430.888916015625, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 1.4328, "grad_norm": 0.06960482895374298, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39575710.0, "reward": 0.957665741443634, "reward_std": 0.41774216294288635, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5090796947479248, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.90234375, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 562.296875, "completions/mean_terminated_length": 455.7500305175781, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.4336, "grad_norm": 0.08485658466815948, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39628449.0, "reward": 0.6886289119720459, "reward_std": 0.34666815400123596, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.36163127422332764, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.796875, "rewards/mean_confidence_reward": 0.6996874809265137, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 650.5625, "completions/mean_terminated_length": 611.9310302734375, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 1.4344000000000001, "grad_norm": 0.031943514943122864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39688021.0, "reward": 1.0000133514404297, "reward_std": 0.35717740654945374, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6093999743461609, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7890625, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 597.671875, "completions/mean_terminated_length": 583.9193115234375, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 1.4352, "grad_norm": 0.04801402613520622, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39742544.0, "reward": 0.8828674554824829, "reward_std": 0.4795517921447754, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4376078248023987, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9045312404632568, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 500.515625, "completions/mean_terminated_length": 483.6290283203125, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.436, "grad_norm": 0.05486871674656868, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39794097.0, "reward": 0.9343204498291016, "reward_std": 0.3210661709308624, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4936390519142151, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8979687690734863, "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": 653.0, "completions/max_terminated_length": 653.0, "completions/mean_length": 407.4375, "completions/mean_terminated_length": 407.4375, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 1.4368, "grad_norm": 0.05816236510872841, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39836077.0, "reward": 0.8556814789772034, "reward_std": 0.35500842332839966, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3988609313964844, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9145312309265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 441.703125, "completions/mean_terminated_length": 432.4603576660156, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.4376, "grad_norm": 0.07894545048475266, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39879962.0, "reward": 0.8348392248153687, "reward_std": 0.41278111934661865, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3884265422821045, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.914843738079071, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 489.0, "completions/mean_terminated_length": 480.5079650878906, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 1.4384000000000001, "grad_norm": 0.05915183946490288, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 39927162.0, "reward": 0.9687556028366089, "reward_std": 0.42663609981536865, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5937593579292297, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7870312929153442, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 504.65625, "completions/mean_terminated_length": 470.0333557128906, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.4392, "grad_norm": 0.07478800415992737, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 39976260.0, "reward": 0.8659180998802185, "reward_std": 0.4454471468925476, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.43495938181877136, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8115624785423279, "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": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 414.5, "completions/mean_terminated_length": 414.5, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 1.44, "grad_norm": 0.078480564057827, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40019380.0, "reward": 1.0014997720718384, "reward_std": 0.3761507272720337, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5342468619346619, "rewards/confidence_one_or_zero": 0.421875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.926562488079071, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 554.65625, "completions/mean_terminated_length": 487.607177734375, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 1.4408, "grad_norm": 0.04771236330270767, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40070702.0, "reward": 0.8164143562316895, "reward_std": 0.34692075848579407, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.414076566696167, "rewards/confidence_one_or_zero": 0.28125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7614046931266785, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 577.984375, "completions/mean_terminated_length": 531.8448486328125, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.4416, "grad_norm": 0.05932563170790672, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40125085.0, "reward": 0.6571970582008362, "reward_std": 0.36555546522140503, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.26751717925071716, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8135937452316284, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 948.0, "completions/mean_length": 509.21875, "completions/mean_terminated_length": 492.6128845214844, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.4424000000000001, "grad_norm": 0.03680983558297157, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40173355.0, "reward": 0.8378681540489197, "reward_std": 0.256045937538147, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.39448437094688416, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9081249833106995, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 799.0, "completions/mean_length": 410.75, "completions/mean_terminated_length": 401.0158996582031, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.4432, "grad_norm": 0.05973440408706665, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40215147.0, "reward": 0.9842798709869385, "reward_std": 0.2664371430873871, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5466827750205994, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8432812690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 583.5, "completions/mean_terminated_length": 569.290283203125, "completions/min_length": 307.0, "completions/min_terminated_length": 307.0, "epoch": 1.444, "grad_norm": 0.07480210065841675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40270059.0, "reward": 0.8026846647262573, "reward_std": 0.39059847593307495, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.37099218368530273, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8935937881469727, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 601.625, "completions/mean_terminated_length": 541.2857666015625, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.4447999999999999, "grad_norm": 0.08735489100217819, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40326675.0, "reward": 0.8712157011032104, "reward_std": 0.519149899482727, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4924297034740448, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.757031261920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 494.0625, "completions/mean_terminated_length": 485.65081787109375, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.4456, "grad_norm": 0.05013938993215561, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40373879.0, "reward": 0.6269846558570862, "reward_std": 0.2510569095611572, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.17584218084812164, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9379687905311584, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 610.453125, "completions/mean_terminated_length": 603.888916015625, "completions/min_length": 304.0, "completions/min_terminated_length": 304.0, "epoch": 1.4464000000000001, "grad_norm": 0.03531401976943016, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40429924.0, "reward": 0.9359346032142639, "reward_std": 0.30261069536209106, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.48124217987060547, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.92578125, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 811.0, "completions/mean_length": 414.015625, "completions/mean_terminated_length": 394.33868408203125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.4472, "grad_norm": 0.08376200497150421, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40473109.0, "reward": 0.9291806221008301, "reward_std": 0.423592209815979, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5146093368530273, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8499999642372131, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 583.421875, "completions/mean_terminated_length": 569.2096557617188, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 1.448, "grad_norm": 0.04265055060386658, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40527040.0, "reward": 0.8017780780792236, "reward_std": 0.35007244348526, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.44730469584465027, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.76171875, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 857.0, "completions/mean_length": 435.21875, "completions/mean_terminated_length": 416.2257995605469, "completions/min_length": 167.0, "completions/min_terminated_length": 167.0, "epoch": 1.4487999999999999, "grad_norm": 0.057235829532146454, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40571422.0, "reward": 0.7232650518417358, "reward_std": 0.23971033096313477, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3059031367301941, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8806250095367432, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 441.328125, "completions/mean_terminated_length": 412.672119140625, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.4496, "grad_norm": 0.04640750586986542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40616419.0, "reward": 1.0404095649719238, "reward_std": 0.2886115610599518, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5808172225952148, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9132812023162842, "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": 511.0, "completions/max_terminated_length": 511.0, "completions/mean_length": 306.5, "completions/mean_terminated_length": 306.5, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 1.4504000000000001, "grad_norm": 0.1109241470694542, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40652755.0, "reward": 1.1122095584869385, "reward_std": 0.4372292757034302, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6306672096252441, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9420312643051147, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 526.953125, "completions/mean_terminated_length": 475.53448486328125, "completions/min_length": 188.0, "completions/min_terminated_length": 188.0, "epoch": 1.4512, "grad_norm": 0.07405867427587509, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40701264.0, "reward": 0.9328978061676025, "reward_std": 0.4628996253013611, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5220437049865723, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8274999856948853, "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": 0.0, "completions/max_length": 1019.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 478.140625, "completions/mean_terminated_length": 478.140625, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.452, "grad_norm": 0.047192033380270004, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40749065.0, "reward": 0.7278783917427063, "reward_std": 0.35622042417526245, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.28387966752052307, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9368749856948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 468.5625, "completions/mean_terminated_length": 459.7460632324219, "completions/min_length": 246.0, "completions/min_terminated_length": 246.0, "epoch": 1.4527999999999999, "grad_norm": 0.04747055470943451, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40794509.0, "reward": 0.8757895231246948, "reward_std": 0.3559808135032654, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4078265428543091, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9464062452316284, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 869.0, "completions/mean_length": 520.78125, "completions/mean_terminated_length": 512.793701171875, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 1.4536, "grad_norm": 0.04570140689611435, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40843791.0, "reward": 1.0869219303131104, "reward_std": 0.32845985889434814, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.626967191696167, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9082655906677246, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 940.0, "completions/mean_length": 569.828125, "completions/mean_terminated_length": 531.3389892578125, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 1.4544000000000001, "grad_norm": 0.07814682275056839, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40896548.0, "reward": 1.0722525119781494, "reward_std": 0.5349303483963013, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.628878116607666, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.878125011920929, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 584.140625, "completions/mean_terminated_length": 569.9515991210938, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 1.4552, "grad_norm": 0.053350262343883514, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 40949933.0, "reward": 0.9794508814811707, "reward_std": 0.43647992610931396, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.552649974822998, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8574999570846558, "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": 941.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 447.078125, "completions/mean_terminated_length": 447.078125, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 1.456, "grad_norm": 0.06091105937957764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 40994466.0, "reward": 0.9634634256362915, "reward_std": 0.29757094383239746, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5206750631332397, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 985.0, "completions/mean_length": 654.65625, "completions/mean_terminated_length": 630.0333862304688, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 1.4567999999999999, "grad_norm": 0.03898937627673149, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41052748.0, "reward": 0.6947765350341797, "reward_std": 0.3374362587928772, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.38955157995224, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.6982812285423279, "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": 0.0, "completions/max_length": 1021.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 385.5625, "completions/mean_terminated_length": 385.5625, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.4576, "grad_norm": 0.033532414585351944, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41092928.0, "reward": 0.926679790019989, "reward_std": 0.23269808292388916, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.49398279190063477, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8920312523841858, "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": 0.0, "completions/max_length": 706.0, "completions/max_terminated_length": 706.0, "completions/mean_length": 380.34375, "completions/mean_terminated_length": 380.34375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.4584, "grad_norm": 0.07771433889865875, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41134262.0, "reward": 0.8981627225875854, "reward_std": 0.394467830657959, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4525734484195709, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9284374713897705, "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": 0.0, "completions/max_length": 452.0, "completions/max_terminated_length": 452.0, "completions/mean_length": 241.53125, "completions/mean_terminated_length": 241.53125, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "epoch": 1.4592, "grad_norm": 0.023556750267744064, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41165128.0, "reward": 1.223895788192749, "reward_std": 0.08568917959928513, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7290390729904175, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9823437929153442, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 447.40625, "completions/mean_terminated_length": 428.8064270019531, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 1.46, "grad_norm": 0.07506537437438965, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41210114.0, "reward": 0.8217071294784546, "reward_std": 0.3431314527988434, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4246624708175659, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8284374475479126, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 589.0, "completions/mean_length": 372.6875, "completions/mean_terminated_length": 362.3492126464844, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.4607999999999999, "grad_norm": 0.06695234775543213, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41249790.0, "reward": 0.8519594073295593, "reward_std": 0.3639543056488037, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.4226672053337097, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7476562857627869, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 861.0, "completions/mean_length": 440.921875, "completions/mean_terminated_length": 422.1128845214844, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.4616, "grad_norm": 0.038163866847753525, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41293849.0, "reward": 0.6224924325942993, "reward_std": 0.26038944721221924, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.18248280882835388, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9114062786102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 474.875, "completions/mean_terminated_length": 466.15875244140625, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.4624, "grad_norm": 0.06208596006035805, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41339633.0, "reward": 0.8189588785171509, "reward_std": 0.33471572399139404, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.35666561126708984, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9409375190734863, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 734.0, "completions/mean_length": 450.75, "completions/mean_terminated_length": 422.5573425292969, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.4632, "grad_norm": 0.06784923374652863, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41385137.0, "reward": 0.8783501386642456, "reward_std": 0.341306209564209, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4441984295845032, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8826375007629395, "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": 603.0, "completions/max_terminated_length": 603.0, "completions/mean_length": 378.265625, "completions/mean_terminated_length": 378.265625, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.464, "grad_norm": 0.05630916357040405, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41425426.0, "reward": 1.2772095203399658, "reward_std": 0.27112025022506714, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.8981671333312988, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.76171875, "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": 599.0, "completions/max_terminated_length": 599.0, "completions/mean_length": 380.859375, "completions/mean_terminated_length": 380.859375, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.4647999999999999, "grad_norm": 0.08981219679117203, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41465193.0, "reward": 1.1318886280059814, "reward_std": 0.41159844398498535, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6543999910354614, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9418749809265137, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 507.53125, "completions/mean_terminated_length": 463.7627258300781, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 1.4656, "grad_norm": 0.058412812650203705, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41514923.0, "reward": 1.0070452690124512, "reward_std": 0.33450472354888916, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6078391075134277, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8064062595367432, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 411.296875, "completions/mean_terminated_length": 401.5714416503906, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.4664, "grad_norm": 0.0616767592728138, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41557294.0, "reward": 1.061058759689331, "reward_std": 0.2877907156944275, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6064906120300293, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9096875190734863, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 444.53125, "completions/mean_terminated_length": 416.03277587890625, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.4672, "grad_norm": 0.05763140693306923, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41602448.0, "reward": 0.9950065016746521, "reward_std": 0.372663289308548, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5681359171867371, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8739062547683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 729.0, "completions/mean_length": 413.5625, "completions/mean_terminated_length": 403.873046875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.468, "grad_norm": 0.056213244795799255, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41644020.0, "reward": 0.8974541425704956, "reward_std": 0.3048246204853058, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.43553122878074646, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9256249666213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 809.0, "completions/mean_length": 458.75, "completions/mean_terminated_length": 449.7778015136719, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 1.4687999999999999, "grad_norm": 0.10480329394340515, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41689844.0, "reward": 0.9111611843109131, "reward_std": 0.3898378014564514, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4629453122615814, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9095312356948853, "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": 672.0, "completions/max_terminated_length": 672.0, "completions/mean_length": 336.859375, "completions/mean_terminated_length": 336.859375, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 1.4696, "grad_norm": 0.14288750290870667, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41726683.0, "reward": 0.9421346783638, "reward_std": 0.3580867052078247, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5405172109603882, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8385937213897705, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 841.0, "completions/mean_length": 401.4375, "completions/mean_terminated_length": 391.5555725097656, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 1.4704, "grad_norm": 0.0716526135802269, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41767591.0, "reward": 0.9824259281158447, "reward_std": 0.44308939576148987, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5117250084877014, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9424999952316284, "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": 817.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 401.625, "completions/mean_terminated_length": 401.625, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 1.4712, "grad_norm": 0.03820282220840454, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41809311.0, "reward": 0.6249626874923706, "reward_std": 0.19832712411880493, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.171798437833786, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9470312595367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 578.5625, "completions/mean_terminated_length": 548.86669921875, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 1.472, "grad_norm": 0.06457985192537308, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 41862179.0, "reward": 0.7364391088485718, "reward_std": 0.3910055458545685, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.34787654876708984, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8148437738418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 846.0, "completions/mean_length": 488.5, "completions/mean_terminated_length": 471.2257995605469, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.4727999999999999, "grad_norm": 0.059786856174468994, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41910851.0, "reward": 0.8437722325325012, "reward_std": 0.351962149143219, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.390667200088501, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9182812571525574, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 551.828125, "completions/mean_terminated_length": 528.6065063476562, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 1.4736, "grad_norm": 0.053613778203725815, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 41961656.0, "reward": 0.7855751514434814, "reward_std": 0.40378350019454956, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.35239842534065247, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8860937356948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 837.0, "completions/mean_length": 383.953125, "completions/mean_terminated_length": 373.7936706542969, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.4744, "grad_norm": 0.07923609018325806, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42002021.0, "reward": 1.1080728769302368, "reward_std": 0.42962926626205444, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6536437273025513, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9118749499320984, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 529.6875, "completions/mean_terminated_length": 505.37701416015625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.4752, "grad_norm": 0.062096498906612396, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42052849.0, "reward": 0.7680572271347046, "reward_std": 0.4288080930709839, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3642374873161316, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8434374928474426, "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": 1007.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 470.09375, "completions/mean_terminated_length": 470.09375, "completions/min_length": 239.0, "completions/min_terminated_length": 239.0, "epoch": 1.476, "grad_norm": 0.04846123605966568, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42099399.0, "reward": 0.9651041030883789, "reward_std": 0.23696979880332947, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4927062690258026, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9543750286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 853.0, "completions/mean_length": 483.0, "completions/mean_terminated_length": 474.4127197265625, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.4768, "grad_norm": 0.04875551164150238, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42147431.0, "reward": 0.9512181282043457, "reward_std": 0.35175377130508423, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5430593490600586, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8474999666213989, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 675.28125, "completions/mean_terminated_length": 602.9056396484375, "completions/min_length": 387.0, "completions/min_terminated_length": 387.0, "epoch": 1.4776, "grad_norm": 0.06133124604821205, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42208361.0, "reward": 0.7206312417984009, "reward_std": 0.38305333256721497, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.4412609338760376, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.5882812738418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 501.765625, "completions/mean_terminated_length": 484.9193420410156, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.4784, "grad_norm": 0.05745622143149376, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42257258.0, "reward": 0.8305580019950867, "reward_std": 0.339054137468338, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4423640966415405, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8248437643051147, "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": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 341.421875, "completions/mean_terminated_length": 341.421875, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.4792, "grad_norm": 0.07662133872509003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42295749.0, "reward": 1.129500389099121, "reward_std": 0.34572088718414307, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.649623453617096, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9567187428474426, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 527.09375, "completions/mean_terminated_length": 519.2063598632812, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 1.48, "grad_norm": 0.04147804528474808, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42345451.0, "reward": 0.851494550704956, "reward_std": 0.2391670048236847, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.4998624920845032, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7676562070846558, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 721.0, "completions/mean_length": 349.453125, "completions/mean_terminated_length": 338.7460632324219, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "epoch": 1.4808, "grad_norm": 0.0753331333398819, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42385224.0, "reward": 0.8749173283576965, "reward_std": 0.3648667633533478, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.546707808971405, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.6645312309265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 452.5, "completions/mean_terminated_length": 443.4285888671875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.4816, "grad_norm": 0.043825745582580566, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42430808.0, "reward": 0.9853276014328003, "reward_std": 0.20979289710521698, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5175281763076782, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9281250238418579, "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": 1019.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 492.4375, "completions/mean_terminated_length": 492.4375, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 1.4824, "grad_norm": 0.07676855474710464, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42477316.0, "reward": 0.9354549646377563, "reward_std": 0.3922930359840393, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5271577835083008, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8592187166213989, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 405.65625, "completions/mean_terminated_length": 395.8412780761719, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.4832, "grad_norm": 0.06487727910280228, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42522734.0, "reward": 0.9627141952514648, "reward_std": 0.3203342854976654, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5191765427589417, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9067187309265137, "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": 817.0, "completions/max_terminated_length": 817.0, "completions/mean_length": 406.203125, "completions/mean_terminated_length": 406.203125, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.484, "grad_norm": 0.07123008370399475, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42565115.0, "reward": 0.7927753329277039, "reward_std": 0.4127359390258789, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.31992343068122864, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9651562571525574, "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": 995.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 499.546875, "completions/mean_terminated_length": 499.546875, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 1.4848, "grad_norm": 0.059334903955459595, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42613454.0, "reward": 0.9195619821548462, "reward_std": 0.4195829927921295, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47974687814712524, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9045312404632568, "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": 0.0, "completions/max_length": 1019.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 560.90625, "completions/mean_terminated_length": 560.90625, "completions/min_length": 335.0, "completions/min_terminated_length": 335.0, "epoch": 1.4856, "grad_norm": 0.05741000920534134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42665000.0, "reward": 0.9239908456802368, "reward_std": 0.3392121195793152, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.45735469460487366, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9398437738418579, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 543.125, "completions/mean_terminated_length": 493.3793029785156, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 1.4864, "grad_norm": 0.08494028449058533, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42716528.0, "reward": 0.6782922148704529, "reward_std": 0.352855920791626, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.3722078204154968, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.6651562452316284, "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": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 508.640625, "completions/mean_terminated_length": 508.640625, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.4872, "grad_norm": 0.07336859405040741, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42765897.0, "reward": 0.8079578876495361, "reward_std": 0.3561168313026428, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.4127890467643738, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8360937833786011, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 464.46875, "completions/mean_terminated_length": 446.4193420410156, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.488, "grad_norm": 0.08258556574583054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42811639.0, "reward": 0.9064385890960693, "reward_std": 0.40216344594955444, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.531624972820282, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7678124904632568, "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": 938.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 412.046875, "completions/mean_terminated_length": 412.046875, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 1.4888, "grad_norm": 0.08903931826353073, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42853546.0, "reward": 0.9438433051109314, "reward_std": 0.29818883538246155, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4814344048500061, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9390625357627869, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 461.46875, "completions/mean_terminated_length": 452.5397033691406, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.4896, "grad_norm": 0.07219576090574265, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 42898648.0, "reward": 1.163960337638855, "reward_std": 0.46729809045791626, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6872937679290771, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9253125190734863, "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": 979.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 438.71875, "completions/mean_terminated_length": 438.71875, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 1.4904, "grad_norm": 0.04715486615896225, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42943686.0, "reward": 1.170069694519043, "reward_std": 0.24917763471603394, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7151374816894531, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9165624976158142, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 436.40625, "completions/mean_terminated_length": 417.45159912109375, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 1.4912, "grad_norm": 0.06366528570652008, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 42989008.0, "reward": 0.8966268301010132, "reward_std": 0.367647647857666, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.44950157403945923, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.90625, "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": 931.0, "completions/max_terminated_length": 931.0, "completions/mean_length": 437.40625, "completions/mean_terminated_length": 437.40625, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.492, "grad_norm": 0.06945546716451645, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43032170.0, "reward": 0.9978721737861633, "reward_std": 0.4369238018989563, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5113672018051147, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9651563167572021, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 876.0, "completions/mean_length": 506.796875, "completions/mean_terminated_length": 498.58734130859375, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 1.4928, "grad_norm": 0.02229131944477558, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43079981.0, "reward": 0.6917883157730103, "reward_std": 0.16548848152160645, "rewards/accuracy_reward": 0.03125, "rewards/brier_reward": 0.3679500222206116, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7381249666213989, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 527.171875, "completions/mean_terminated_length": 485.06781005859375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 1.4936, "grad_norm": 0.04418392479419708, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43129976.0, "reward": 0.5487173795700073, "reward_std": 0.17061716318130493, "rewards/accuracy_reward": 0.015625, "rewards/brier_reward": 0.15993282198905945, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8226562142372131, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 979.0, "completions/mean_length": 577.0625, "completions/mean_terminated_length": 555.0819091796875, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 1.4944, "grad_norm": 0.05017677694559097, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43183820.0, "reward": 0.8238016366958618, "reward_std": 0.4055600166320801, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.42885154485702515, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8404687643051147, "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": 952.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 342.078125, "completions/mean_terminated_length": 342.078125, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 1.4952, "grad_norm": 0.05068056657910347, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43222673.0, "reward": 0.9503347277641296, "reward_std": 0.16616089642047882, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4944171905517578, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9320312738418579, "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": 0.1875, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 697.625, "completions/mean_terminated_length": 622.3077392578125, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 1.496, "grad_norm": 0.05816454812884331, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43283145.0, "reward": 0.8233202695846558, "reward_std": 0.4448412358760834, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.49038904905319214, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.796875, "rewards/mean_confidence_reward": 0.6814062595367432, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 502.765625, "completions/mean_terminated_length": 448.8448181152344, "completions/min_length": 131.0, "completions/min_terminated_length": 131.0, "epoch": 1.4968, "grad_norm": 0.0684436485171318, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43332874.0, "reward": 0.9056978225708008, "reward_std": 0.3379165530204773, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5457687377929688, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7662500143051147, "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": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 417.484375, "completions/mean_terminated_length": 417.484375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.4976, "grad_norm": 0.04962118715047836, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43375385.0, "reward": 0.894334614276886, "reward_std": 0.2529928982257843, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4292921721935272, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9420312643051147, "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": 636.0, "completions/max_terminated_length": 636.0, "completions/mean_length": 312.75, "completions/mean_terminated_length": 312.75, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.4984, "grad_norm": 0.055982038378715515, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43411513.0, "reward": 1.3017714023590088, "reward_std": 0.3214719295501709, "rewards/accuracy_reward": 0.78125, "rewards/brier_reward": 0.8222906589508057, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9462499618530273, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 790.0, "completions/mean_length": 449.921875, "completions/mean_terminated_length": 411.6500244140625, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 1.4992, "grad_norm": 0.06807852536439896, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43456388.0, "reward": 1.1290953159332275, "reward_std": 0.3949526250362396, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.7113140225410461, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8348437547683716, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 935.0, "completions/mean_length": 564.859375, "completions/mean_terminated_length": 542.2786865234375, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 1.5, "grad_norm": 0.055951084941625595, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43508523.0, "reward": 0.9748731255531311, "reward_std": 0.4833749830722809, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5122437477111816, "rewards/confidence_one_or_zero": 0.328125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.925000011920929, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 510.171875, "completions/mean_terminated_length": 484.901611328125, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.5008, "grad_norm": 0.07863949984312057, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43557766.0, "reward": 0.8708556890487671, "reward_std": 0.45565614104270935, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4448343813419342, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8876562118530273, "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": 728.0, "completions/max_terminated_length": 728.0, "completions/mean_length": 321.546875, "completions/mean_terminated_length": 321.546875, "completions/min_length": 104.0, "completions/min_terminated_length": 104.0, "epoch": 1.5016, "grad_norm": 0.046180304139852524, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43593641.0, "reward": 0.8406533002853394, "reward_std": 0.20518147945404053, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3688046932220459, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9560922384262085, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 915.0, "completions/mean_length": 543.75, "completions/mean_terminated_length": 536.1270141601562, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 1.5024, "grad_norm": 0.026315297931432724, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43644313.0, "reward": 0.7963610887527466, "reward_std": 0.1737842857837677, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.38959529995918274, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.867968738079071, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 488.09375, "completions/mean_terminated_length": 479.58734130859375, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.5032, "grad_norm": 0.060063935816287994, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43691935.0, "reward": 0.8911619782447815, "reward_std": 0.2950814962387085, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.46982187032699585, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8742187023162842, "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": 0.0, "completions/max_length": 781.0, "completions/max_terminated_length": 781.0, "completions/mean_length": 457.421875, "completions/mean_terminated_length": 457.421875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.504, "grad_norm": 0.07690071314573288, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43737114.0, "reward": 0.9818009734153748, "reward_std": 0.4459003508090973, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5261000394821167, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9209374785423279, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1005.0, "completions/mean_length": 428.78125, "completions/mean_terminated_length": 389.1000061035156, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.5048, "grad_norm": 0.08015302568674088, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43779852.0, "reward": 0.8533720374107361, "reward_std": 0.5119991302490234, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.42549219727516174, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8682812452316284, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 550.421875, "completions/mean_terminated_length": 535.1451416015625, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 1.5056, "grad_norm": 0.03326240926980972, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43830903.0, "reward": 0.7046048641204834, "reward_std": 0.28755074739456177, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.3310827910900116, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8035781383514404, "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": 1008.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 573.390625, "completions/mean_terminated_length": 573.390625, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 1.5064, "grad_norm": 0.035733163356781006, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 43884880.0, "reward": 1.008389949798584, "reward_std": 0.3365405201911926, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5792781114578247, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8759374618530273, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 914.0, "completions/mean_length": 428.125, "completions/mean_terminated_length": 408.9032287597656, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 1.5072, "grad_norm": 0.08784973621368408, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43929464.0, "reward": 0.7963690161705017, "reward_std": 0.2947612404823303, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.3427359461784363, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9334375262260437, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 733.421875, "completions/mean_terminated_length": 697.7368774414062, "completions/min_length": 497.0, "completions/min_terminated_length": 497.0, "epoch": 1.508, "grad_norm": 0.05038806051015854, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 43992755.0, "reward": 0.6599618196487427, "reward_std": 0.3600037395954132, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.3355468809604645, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7296874523162842, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 589.25, "completions/mean_terminated_length": 560.2667236328125, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.5088, "grad_norm": 0.032203592360019684, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44046899.0, "reward": 0.9042148590087891, "reward_std": 0.331127792596817, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5428031086921692, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7704687714576721, "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": 950.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 425.78125, "completions/mean_terminated_length": 425.78125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 1.5096, "grad_norm": 0.05351550877094269, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44089925.0, "reward": 0.9517072439193726, "reward_std": 0.3181668221950531, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5596625208854675, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8087500333786011, "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": 0.0, "completions/max_length": 782.0, "completions/max_terminated_length": 782.0, "completions/mean_length": 487.21875, "completions/mean_terminated_length": 487.21875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.5104, "grad_norm": 0.06958874315023422, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44137731.0, "reward": 0.9404415488243103, "reward_std": 0.3345091938972473, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4746311902999878, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9546562433242798, "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": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 380.46875, "completions/mean_terminated_length": 380.46875, "completions/min_length": 135.0, "completions/min_terminated_length": 135.0, "epoch": 1.5112, "grad_norm": 0.048459332436323166, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44177745.0, "reward": 1.123884677886963, "reward_std": 0.20472495257854462, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6383922100067139, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9645313024520874, "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": 0.0, "completions/max_length": 762.0, "completions/max_terminated_length": 762.0, "completions/mean_length": 434.390625, "completions/mean_terminated_length": 434.390625, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 1.512, "grad_norm": 0.04331769421696663, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44222538.0, "reward": 1.3046133518218994, "reward_std": 0.34423041343688965, "rewards/accuracy_reward": 0.796875, "rewards/brier_reward": 0.8123500347137451, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.956250011920929, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 536.765625, "completions/mean_terminated_length": 512.80322265625, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 1.5128, "grad_norm": 0.04768502339720726, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44274363.0, "reward": 0.9947397708892822, "reward_std": 0.3458603024482727, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.6769781112670898, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.643750011920929, "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": 928.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 443.78125, "completions/mean_terminated_length": 443.78125, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.5135999999999998, "grad_norm": 0.06391414254903793, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44318141.0, "reward": 1.098233938217163, "reward_std": 0.39461246132850647, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6339656114578247, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9365624785423279, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 535.171875, "completions/mean_terminated_length": 455.1817932128906, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 1.5144, "grad_norm": 0.06807396560907364, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44368536.0, "reward": 0.8816843628883362, "reward_std": 0.47815433144569397, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4977422058582306, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.7946875095367432, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 551.4375, "completions/mean_terminated_length": 502.5517272949219, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 1.5152, "grad_norm": 0.04333198815584183, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44421060.0, "reward": 0.9421048164367676, "reward_std": 0.3737607002258301, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5404577851295471, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8246874809265137, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 847.0, "completions/mean_length": 440.21875, "completions/mean_terminated_length": 430.952392578125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 1.516, "grad_norm": 0.09606114029884338, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44464642.0, "reward": 0.7960472106933594, "reward_std": 0.3982251286506653, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3733421862125397, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8707813024520874, "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": 0.0, "completions/max_length": 881.0, "completions/max_terminated_length": 881.0, "completions/mean_length": 417.421875, "completions/mean_terminated_length": 417.421875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 1.5168, "grad_norm": 0.06068497151136398, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44508429.0, "reward": 0.9057705402374268, "reward_std": 0.3885166049003601, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.45216408371925354, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9126561880111694, "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": 723.0, "completions/max_terminated_length": 723.0, "completions/mean_length": 374.453125, "completions/mean_terminated_length": 374.453125, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 1.5175999999999998, "grad_norm": 0.04993486776947975, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44549162.0, "reward": 1.1196937561035156, "reward_std": 0.35082298517227173, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.7393859624862671, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.766406238079071, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 575.34375, "completions/mean_terminated_length": 482.2264099121094, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.5184, "grad_norm": 0.05254688113927841, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44600848.0, "reward": 0.5941022634506226, "reward_std": 0.3295300602912903, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.3132030963897705, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.796875, "rewards/mean_confidence_reward": 0.653124988079071, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 437.0625, "completions/mean_terminated_length": 418.1290283203125, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.5192, "grad_norm": 0.062067024409770966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44645188.0, "reward": 1.1257274150848389, "reward_std": 0.35124337673187256, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6733281016349792, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9031250476837158, "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": 761.0, "completions/max_terminated_length": 761.0, "completions/mean_length": 415.5625, "completions/mean_terminated_length": 415.5625, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.52, "grad_norm": 0.06874002516269684, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44687832.0, "reward": 1.0964345932006836, "reward_std": 0.46120911836624146, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6303671598434448, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9329687356948853, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 504.9375, "completions/mean_terminated_length": 420.0, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.5208, "grad_norm": 0.06764552742242813, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44735204.0, "reward": 0.9839797019958496, "reward_std": 0.4071544408798218, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5929578542709351, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7795312404632568, "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": 981.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 552.734375, "completions/mean_terminated_length": 552.734375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 1.5215999999999998, "grad_norm": 0.05160878226161003, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44788451.0, "reward": 0.921068012714386, "reward_std": 0.34440138936042786, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.5608843564987183, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.770312488079071, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 944.0, "completions/mean_length": 456.75, "completions/mean_terminated_length": 428.8524475097656, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 1.5224, "grad_norm": 0.04870004579424858, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44834963.0, "reward": 0.7277376651763916, "reward_std": 0.32669365406036377, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3304734230041504, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8314062356948853, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 557.125, "completions/mean_terminated_length": 517.559326171875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.5232, "grad_norm": 0.061917487531900406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44889915.0, "reward": 0.6964244842529297, "reward_std": 0.34627804160118103, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.29909688234329224, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8143749833106995, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 553.765625, "completions/mean_terminated_length": 530.6393432617188, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 1.524, "grad_norm": 0.05110371485352516, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 44941212.0, "reward": 0.9812774658203125, "reward_std": 0.4126396179199219, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5406781435012817, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8799999952316284, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 529.875, "completions/mean_terminated_length": 449.0181579589844, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 1.5248, "grad_norm": 0.06520368903875351, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 44991396.0, "reward": 1.0030499696731567, "reward_std": 0.3952364921569824, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5998484492301941, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7985937595367432, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 970.0, "completions/mean_length": 464.34375, "completions/mean_terminated_length": 446.2903137207031, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.5255999999999998, "grad_norm": 0.05511585250496864, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45037722.0, "reward": 0.9641024470329285, "reward_std": 0.3335370719432831, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5688281059265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8226562738418579, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1017.0, "completions/mean_length": 644.421875, "completions/mean_terminated_length": 619.11669921875, "completions/min_length": 339.0, "completions/min_terminated_length": 339.0, "epoch": 1.5264, "grad_norm": 0.028827277943491936, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45095525.0, "reward": 0.6257204413414001, "reward_std": 0.26463672518730164, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.26706406474113464, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7903125286102295, "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": 883.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 391.234375, "completions/mean_terminated_length": 391.234375, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 1.5272000000000001, "grad_norm": 0.09372863918542862, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45136148.0, "reward": 0.7525534629821777, "reward_std": 0.3341353237628937, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.301979660987854, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9326562285423279, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 691.0, "completions/mean_length": 396.359375, "completions/mean_terminated_length": 386.3968505859375, "completions/min_length": 230.0, "completions/min_terminated_length": 230.0, "epoch": 1.528, "grad_norm": 0.06386056542396545, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45176283.0, "reward": 0.7523165941238403, "reward_std": 0.3757275640964508, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.33275628089904785, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8856250047683716, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 662.5, "completions/mean_terminated_length": 638.4000244140625, "completions/min_length": 364.0, "completions/min_terminated_length": 364.0, "epoch": 1.5288, "grad_norm": 0.04231744259595871, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45235883.0, "reward": 0.6704657077789307, "reward_std": 0.34301134943962097, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.29405468702316284, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7920312881469727, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 884.0, "completions/mean_length": 583.671875, "completions/mean_terminated_length": 554.3167114257812, "completions/min_length": 315.0, "completions/min_terminated_length": 315.0, "epoch": 1.5295999999999998, "grad_norm": 0.04006511718034744, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45289814.0, "reward": 0.9077163934707642, "reward_std": 0.36079490184783936, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5341812372207642, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7853125333786011, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 605.421875, "completions/mean_terminated_length": 562.1206665039062, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 1.5304, "grad_norm": 0.07221516966819763, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45345969.0, "reward": 0.786638617515564, "reward_std": 0.4976189434528351, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.38577499985694885, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7978124618530273, "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": 720.0, "completions/max_terminated_length": 720.0, "completions/mean_length": 367.1875, "completions/mean_terminated_length": 367.1875, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 1.5312000000000001, "grad_norm": 0.06141427159309387, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45386269.0, "reward": 1.1658048629760742, "reward_std": 0.2438219040632248, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6909828186035156, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9323437213897705, "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": 770.0, "completions/max_terminated_length": 770.0, "completions/mean_length": 414.4375, "completions/mean_terminated_length": 414.4375, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.532, "grad_norm": 0.04298411309719086, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45429801.0, "reward": 1.092833161354065, "reward_std": 0.2377612590789795, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6231640577316284, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.94140625, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 572.515625, "completions/mean_terminated_length": 498.6363525390625, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 1.5328, "grad_norm": 0.050191063433885574, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45482266.0, "reward": 1.097325086593628, "reward_std": 0.3670763373374939, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.7102733850479126, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7654375433921814, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 542.96875, "completions/mean_terminated_length": 510.9000244140625, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 1.5335999999999999, "grad_norm": 0.043093543499708176, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45533576.0, "reward": 0.962704062461853, "reward_std": 0.4181201159954071, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5347812175750732, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8603125214576721, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 738.828125, "completions/mean_terminated_length": 686.0184936523438, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.5344, "grad_norm": 0.0370011068880558, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45599341.0, "reward": 0.8482953310012817, "reward_std": 0.4708915948867798, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5715892314910889, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.5940060615539551, "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": 974.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 450.703125, "completions/mean_terminated_length": 450.703125, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.5352000000000001, "grad_norm": 0.07563259452581406, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45643242.0, "reward": 1.1138982772827148, "reward_std": 0.3369200825691223, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.7277949452400208, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8241406083106995, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 578.828125, "completions/mean_terminated_length": 524.1578979492188, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 1.536, "grad_norm": 0.038350269198417664, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45697487.0, "reward": 0.9250626564025879, "reward_std": 0.3446233868598938, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.553248405456543, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7890625, "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": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 441.4375, "completions/mean_terminated_length": 441.4375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 1.5368, "grad_norm": 0.0707312524318695, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45741195.0, "reward": 0.9759385585784912, "reward_std": 0.39145517349243164, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5299999117851257, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9162187576293945, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 509.609375, "completions/mean_terminated_length": 501.4444885253906, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.5375999999999999, "grad_norm": 0.033173058182001114, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45790930.0, "reward": 0.9298768043518066, "reward_std": 0.2927643954753876, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5160015821456909, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8567187786102295, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 671.6875, "completions/mean_terminated_length": 614.0363159179688, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.5384, "grad_norm": 0.058211129158735275, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45850174.0, "reward": 0.6646883487701416, "reward_std": 0.37982600927352905, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.2668750286102295, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8171874284744263, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 522.828125, "completions/mean_terminated_length": 480.3559265136719, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 1.5392000000000001, "grad_norm": 0.05754833295941353, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45899139.0, "reward": 1.0127549171447754, "reward_std": 0.399946928024292, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6192578077316284, "rewards/confidence_one_or_zero": 0.34375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.6526562571525574, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 665.0, "completions/mean_length": 433.875, "completions/mean_terminated_length": 424.5079650878906, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.54, "grad_norm": 0.0817500576376915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 45944219.0, "reward": 0.785362720489502, "reward_std": 0.4299655258655548, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.32072341442108154, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.946718692779541, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 582.671875, "completions/mean_terminated_length": 545.2711791992188, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 1.5408, "grad_norm": 0.06466224044561386, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 45997574.0, "reward": 0.7985594868659973, "reward_std": 0.3990691304206848, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.37836718559265137, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8617187738418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 596.75, "completions/mean_terminated_length": 582.9677124023438, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 1.5415999999999999, "grad_norm": 0.04326237365603447, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46051654.0, "reward": 0.8898774981498718, "reward_std": 0.36441561579704285, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.4672531187534332, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8806250095367432, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 910.0, "completions/mean_length": 620.3125, "completions/mean_terminated_length": 536.5283203125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.5424, "grad_norm": 0.042164143174886703, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46107274.0, "reward": 0.6421008706092834, "reward_std": 0.3594595193862915, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.3154500126838684, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.8125, "rewards/mean_confidence_reward": 0.7065625190734863, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 568.5625, "completions/mean_terminated_length": 521.4483032226562, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.5432000000000001, "grad_norm": 0.0344642736017704, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46159790.0, "reward": 0.7198602557182312, "reward_std": 0.27433305978775024, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.29909375309944153, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8546874523162842, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 562.640625, "completions/mean_terminated_length": 539.9508056640625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 1.544, "grad_norm": 0.05542374774813652, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46211671.0, "reward": 1.0447250604629517, "reward_std": 0.4368487000465393, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6363234519958496, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8306249380111694, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 624.703125, "completions/mean_terminated_length": 590.8643798828125, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 1.5448, "grad_norm": 0.05640164762735367, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46268100.0, "reward": 0.7727047801017761, "reward_std": 0.42378702759742737, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.46728280186653137, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6714062690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 457.3125, "completions/mean_terminated_length": 448.3174743652344, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 1.5455999999999999, "grad_norm": 0.049102071672677994, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46312920.0, "reward": 0.854557991027832, "reward_std": 0.33781397342681885, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.45911407470703125, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7923437356948853, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 618.609375, "completions/mean_terminated_length": 568.8245849609375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.5464, "grad_norm": 0.06803382933139801, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46368287.0, "reward": 0.7796226739883423, "reward_std": 0.4262527823448181, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.48111873865127563, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.645312488079071, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 436.25, "completions/mean_terminated_length": 426.920654296875, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 1.5472000000000001, "grad_norm": 0.0780922919511795, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46412079.0, "reward": 1.1332931518554688, "reward_std": 0.3859596252441406, "rewards/accuracy_reward": 0.625, "rewards/brier_reward": 0.6572093963623047, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8278124928474426, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 592.609375, "completions/mean_terminated_length": 539.631591796875, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 1.548, "grad_norm": 0.058212872594594955, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46465638.0, "reward": 0.9043289422988892, "reward_std": 0.2970638573169708, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5742812156677246, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.6878124475479126, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 676.484375, "completions/mean_terminated_length": 659.3934326171875, "completions/min_length": 428.0, "completions/min_terminated_length": 428.0, "epoch": 1.5488, "grad_norm": 0.03549567982554436, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46526837.0, "reward": 0.7419710755348206, "reward_std": 0.2642650306224823, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.46831560134887695, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6584374904632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 363.0, "completions/mean_terminated_length": 352.5079650878906, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 1.5495999999999999, "grad_norm": 0.04633718729019165, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46565541.0, "reward": 1.0941877365112305, "reward_std": 0.17183610796928406, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6414984464645386, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8882812261581421, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 476.15625, "completions/mean_terminated_length": 439.63336181640625, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 1.5504, "grad_norm": 0.044924043118953705, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46611135.0, "reward": 1.1346439123153687, "reward_std": 0.363506555557251, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6755359172821045, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8942187428474426, "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": 917.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 503.28125, "completions/mean_terminated_length": 503.28125, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 1.5512000000000001, "grad_norm": 0.06059780344367027, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46658481.0, "reward": 0.9768993854522705, "reward_std": 0.3185531198978424, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5162968635559082, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9340624809265137, "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": 0.0, "completions/max_length": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 476.671875, "completions/mean_terminated_length": 476.671875, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.552, "grad_norm": 0.05387941002845764, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46704972.0, "reward": 0.9244095087051392, "reward_std": 0.30737632513046265, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.5206922292709351, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8526562452316284, "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": 0.0, "completions/max_length": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 354.671875, "completions/mean_terminated_length": 354.671875, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 1.5528, "grad_norm": 0.08216128498315811, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46743223.0, "reward": 1.1047087907791138, "reward_std": 0.43363022804260254, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6156656742095947, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9521875381469727, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 664.5625, "completions/mean_terminated_length": 605.7454223632812, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 1.5535999999999999, "grad_norm": 0.057383764535188675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46804139.0, "reward": 0.7978687286376953, "reward_std": 0.4177854061126709, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.48636093735694885, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.6676562428474426, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 713.0, "completions/mean_length": 477.671875, "completions/mean_terminated_length": 469.0000305175781, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.5544, "grad_norm": 0.04036366194486618, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46852598.0, "reward": 0.7748492956161499, "reward_std": 0.33156195282936096, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3778218924999237, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7678124904632568, "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": 690.0, "completions/max_terminated_length": 690.0, "completions/mean_length": 383.828125, "completions/mean_terminated_length": 383.828125, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.5552000000000001, "grad_norm": 0.05677270516753197, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46892891.0, "reward": 1.0385322570800781, "reward_std": 0.3622620701789856, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5614374876022339, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9565625190734863, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 633.359375, "completions/mean_terminated_length": 585.385986328125, "completions/min_length": 297.0, "completions/min_terminated_length": 297.0, "epoch": 1.556, "grad_norm": 0.057494811713695526, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 46949810.0, "reward": 0.843582808971405, "reward_std": 0.4584164023399353, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4684140682220459, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7735937833786011, "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": 737.0, "completions/max_terminated_length": 737.0, "completions/mean_length": 422.34375, "completions/mean_terminated_length": 422.34375, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 1.5568, "grad_norm": 0.07217943668365479, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 46992984.0, "reward": 1.3003370761871338, "reward_std": 0.3380177617073059, "rewards/accuracy_reward": 0.796875, "rewards/brier_reward": 0.8194218873977661, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9535937309265137, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 594.96875, "completions/mean_terminated_length": 566.36669921875, "completions/min_length": 327.0, "completions/min_terminated_length": 327.0, "epoch": 1.5575999999999999, "grad_norm": 0.05136286839842796, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47047238.0, "reward": 0.9256938695907593, "reward_std": 0.3993191421031952, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5388859510421753, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7926562428474426, "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": 771.0, "completions/max_terminated_length": 771.0, "completions/mean_length": 378.203125, "completions/mean_terminated_length": 378.203125, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.5584, "grad_norm": 0.05044078454375267, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47087347.0, "reward": 1.235190749168396, "reward_std": 0.27008137106895447, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.7672547101974487, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9260937571525574, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 856.0, "completions/mean_length": 449.5625, "completions/mean_terminated_length": 421.31146240234375, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.5592000000000001, "grad_norm": 0.07715071737766266, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47132039.0, "reward": 1.1145284175872803, "reward_std": 0.4367232620716095, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6665546894073486, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8870312571525574, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 974.0, "completions/mean_length": 543.71875, "completions/mean_terminated_length": 528.2257690429688, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 1.56, "grad_norm": 0.04822012037038803, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47184325.0, "reward": 0.7991681098937988, "reward_std": 0.4000406265258789, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3639593720436096, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8909374475479126, "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": 795.0, "completions/max_terminated_length": 795.0, "completions/mean_length": 421.59375, "completions/mean_terminated_length": 421.59375, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 1.5608, "grad_norm": 0.06784866750240326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47226379.0, "reward": 1.0621581077575684, "reward_std": 0.3762356638908386, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6555640697479248, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8167186975479126, "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": 778.0, "completions/max_terminated_length": 778.0, "completions/mean_length": 414.6875, "completions/mean_terminated_length": 414.6875, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.5615999999999999, "grad_norm": 0.07531047612428665, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47268599.0, "reward": 0.7478829622268677, "reward_std": 0.33549290895462036, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3238890469074249, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8848437666893005, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 441.0625, "completions/mean_terminated_length": 431.8095397949219, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 1.5624, "grad_norm": 0.07893100380897522, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47313643.0, "reward": 0.9410722255706787, "reward_std": 0.3741060197353363, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.47589218616485596, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9292187690734863, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 845.0, "completions/mean_length": 401.859375, "completions/mean_terminated_length": 391.9841613769531, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 1.5632000000000001, "grad_norm": 0.03918377310037613, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47356930.0, "reward": 0.6367549896240234, "reward_std": 0.2634134590625763, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.22663280367851257, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8601562976837158, "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": 649.0, "completions/max_terminated_length": 649.0, "completions/mean_length": 363.71875, "completions/mean_terminated_length": 363.71875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.564, "grad_norm": 0.07174400985240936, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47396560.0, "reward": 1.0669556856155396, "reward_std": 0.3299671411514282, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5870343446731567, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9596874713897705, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 564.375, "completions/mean_terminated_length": 533.7333374023438, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.5648, "grad_norm": 0.04124074801802635, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47450168.0, "reward": 1.1252212524414062, "reward_std": 0.4017171263694763, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.7035655975341797, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8503124713897705, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 814.0, "completions/mean_length": 527.890625, "completions/mean_terminated_length": 511.8870849609375, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 1.5655999999999999, "grad_norm": 0.044096242636442184, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47501249.0, "reward": 0.9225736260414124, "reward_std": 0.3977367877960205, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4857703447341919, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8804531097412109, "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": 864.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 493.515625, "completions/mean_terminated_length": 493.515625, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.5664, "grad_norm": 0.0766482725739479, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47549074.0, "reward": 0.9161728620529175, "reward_std": 0.4466990828514099, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4573437571525574, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.932812511920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 894.0, "completions/mean_length": 471.234375, "completions/mean_terminated_length": 462.4603576660156, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.5672000000000001, "grad_norm": 0.0574045330286026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47595905.0, "reward": 0.9769110679626465, "reward_std": 0.4138144254684448, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5163203477859497, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8979687094688416, "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": 892.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 519.0625, "completions/mean_terminated_length": 519.0625, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 1.568, "grad_norm": 0.05875152722001076, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47645173.0, "reward": 0.8754806518554688, "reward_std": 0.4188961386680603, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.4697093963623047, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8543750047683716, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 514.59375, "completions/mean_terminated_length": 506.5079650878906, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.5688, "grad_norm": 0.056376248598098755, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47692779.0, "reward": 0.7594572305679321, "reward_std": 0.2715509831905365, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.31578749418258667, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9225000143051147, "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": 0.0, "completions/max_length": 740.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 398.3125, "completions/mean_terminated_length": 398.3125, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.5695999999999999, "grad_norm": 0.10510324686765671, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47733983.0, "reward": 1.0061259269714355, "reward_std": 0.3415524959564209, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5435000061988831, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9401562213897705, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 505.234375, "completions/mean_terminated_length": 441.52630615234375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 1.5704, "grad_norm": 0.042174119502305984, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47783358.0, "reward": 0.8225687742233276, "reward_std": 0.2856256365776062, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.42638593912124634, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8114062547683716, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 474.09375, "completions/mean_terminated_length": 447.0491638183594, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 1.5712000000000002, "grad_norm": 0.049471937119960785, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47829124.0, "reward": 1.0360502004623413, "reward_std": 0.2504999041557312, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.603348433971405, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8851562738418579, "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": 992.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 325.703125, "completions/mean_terminated_length": 325.703125, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.572, "grad_norm": 0.07780090719461441, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 47865921.0, "reward": 1.2470464706420898, "reward_std": 0.4217509627342224, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7753406167030334, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9299999475479126, "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": 983.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 352.859375, "completions/mean_terminated_length": 352.859375, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.5728, "grad_norm": 0.07521709054708481, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47904088.0, "reward": 1.1704862117767334, "reward_std": 0.2660471796989441, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6847203373908997, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9682812690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 924.0, "completions/mean_length": 588.625, "completions/mean_terminated_length": 574.5806274414062, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 1.5735999999999999, "grad_norm": 0.03533731773495674, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 47958160.0, "reward": 0.7014329433441162, "reward_std": 0.20747488737106323, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.2934890389442444, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8751562833786011, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 526.390625, "completions/mean_terminated_length": 510.33868408203125, "completions/min_length": 277.0, "completions/min_terminated_length": 277.0, "epoch": 1.5744, "grad_norm": 0.05323140323162079, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48008505.0, "reward": 0.9318039417266846, "reward_std": 0.41640618443489075, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5511062741279602, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7868750095367432, "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": 0.0, "completions/max_length": 745.0, "completions/max_terminated_length": 745.0, "completions/mean_length": 360.46875, "completions/mean_terminated_length": 360.46875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.5752000000000002, "grad_norm": 0.05347572639584541, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48049511.0, "reward": 1.3475931882858276, "reward_std": 0.2876591384410858, "rewards/accuracy_reward": 0.828125, "rewards/brier_reward": 0.8670593500137329, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9359375238418579, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 618.234375, "completions/mean_terminated_length": 568.4035034179688, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 1.576, "grad_norm": 0.05780630558729172, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48107702.0, "reward": 0.7842226624488831, "reward_std": 0.4967570900917053, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.41219374537467957, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7565624713897705, "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": 977.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 453.03125, "completions/mean_terminated_length": 453.03125, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 1.5768, "grad_norm": 0.04664647579193115, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48153992.0, "reward": 1.0393697023391724, "reward_std": 0.3271281123161316, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6412374973297119, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8165625333786011, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 559.28125, "completions/mean_terminated_length": 551.90478515625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.5776, "grad_norm": 0.04232216253876686, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48208938.0, "reward": 0.809107780456543, "reward_std": 0.2449510544538498, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.49321404099464417, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7179688215255737, "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": 885.0, "completions/max_terminated_length": 885.0, "completions/mean_length": 486.4375, "completions/mean_terminated_length": 486.4375, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 1.5784, "grad_norm": 0.07329200953245163, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48256166.0, "reward": 0.9437837600708008, "reward_std": 0.4038074016571045, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5125656127929688, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8909374475479126, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 993.0, "completions/mean_length": 534.265625, "completions/mean_terminated_length": 518.4677124023438, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.5792000000000002, "grad_norm": 0.052561987191438675, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48309127.0, "reward": 1.1429274082183838, "reward_std": 0.36293596029281616, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.7858531475067139, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7306250333786011, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 517.71875, "completions/mean_terminated_length": 509.68255615234375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 1.58, "grad_norm": 0.05821007862687111, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48359701.0, "reward": 1.1231814622879028, "reward_std": 0.4470527172088623, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6526108980178833, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9360937476158142, "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": 0.328125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1023.0, "completions/mean_length": 889.703125, "completions/mean_terminated_length": 824.1162719726562, "completions/min_length": 494.0, "completions/min_terminated_length": 494.0, "epoch": 1.5808, "grad_norm": 0.03996993228793144, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48434530.0, "reward": 0.563038170337677, "reward_std": 0.48059776425361633, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.31357499957084656, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.65625, "rewards/mean_confidence_reward": 0.5362499952316284, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 581.96875, "completions/mean_terminated_length": 567.7096557617188, "completions/min_length": 336.0, "completions/min_terminated_length": 336.0, "epoch": 1.5816, "grad_norm": 0.04404616355895996, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48489168.0, "reward": 0.9079493284225464, "reward_std": 0.40589383244514465, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5346468687057495, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7743749618530273, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 903.0, "completions/mean_length": 448.5625, "completions/mean_terminated_length": 439.4285888671875, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 1.5824, "grad_norm": 0.08302208036184311, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48533620.0, "reward": 1.097775936126709, "reward_std": 0.4021092355251312, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6799249649047852, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8421875238418579, "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": 880.0, "completions/mean_length": 574.765625, "completions/mean_terminated_length": 544.8167114257812, "completions/min_length": 278.0, "completions/min_terminated_length": 278.0, "epoch": 1.5832000000000002, "grad_norm": 0.04143305867910385, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48588389.0, "reward": 0.9706423282623291, "reward_std": 0.22443611919879913, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5350328087806702, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8723437786102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 575.125, "completions/mean_terminated_length": 545.2000122070312, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.584, "grad_norm": 0.0814661830663681, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48641277.0, "reward": 0.8923267126083374, "reward_std": 0.443517804145813, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47215157747268677, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8660937547683716, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 564.515625, "completions/mean_terminated_length": 533.8833618164062, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 1.5848, "grad_norm": 0.04812004044651985, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48693694.0, "reward": 1.002110242843628, "reward_std": 0.3110150098800659, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.7229687571525574, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.596875011920929, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 896.0, "completions/mean_length": 625.578125, "completions/mean_terminated_length": 591.8135375976562, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.5856, "grad_norm": 0.07139583677053452, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48749763.0, "reward": 0.8986057043075562, "reward_std": 0.3675723671913147, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.46908438205718994, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8590624332427979, "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": 843.0, "completions/max_terminated_length": 843.0, "completions/mean_length": 395.015625, "completions/mean_terminated_length": 395.015625, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.5864, "grad_norm": 0.04897815361618996, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48792484.0, "reward": 0.7254495620727539, "reward_std": 0.2942624092102051, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.26339685916900635, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9506250023841858, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 972.0, "completions/mean_length": 613.25, "completions/mean_terminated_length": 537.1851806640625, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.5872000000000002, "grad_norm": 0.057491712272167206, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48848308.0, "reward": 0.9244673252105713, "reward_std": 0.4350977838039398, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5208078026771545, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.8045312762260437, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 455.296875, "completions/mean_terminated_length": 446.2698669433594, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.588, "grad_norm": 0.06633297353982925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48893319.0, "reward": 0.9615814685821533, "reward_std": 0.3163769245147705, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5169109106063843, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9004687666893005, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 858.0, "completions/mean_length": 488.375, "completions/mean_terminated_length": 479.873046875, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.5888, "grad_norm": 0.05881144851446152, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 48940735.0, "reward": 1.0475047826766968, "reward_std": 0.3965556025505066, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.5481328368186951, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7407812476158142, "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": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 462.09375, "completions/mean_terminated_length": 462.09375, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.5896, "grad_norm": 0.055717263370752335, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 48986965.0, "reward": 1.0065563917160034, "reward_std": 0.3556148111820221, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5599859356880188, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8707812428474426, "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": 868.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 404.0625, "completions/mean_terminated_length": 404.0625, "completions/min_length": 233.0, "completions/min_terminated_length": 233.0, "epoch": 1.5904, "grad_norm": 0.06122602894902229, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49029321.0, "reward": 1.1375250816345215, "reward_std": 0.27571678161621094, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.7594234347343445, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.764843761920929, "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": 990.0, "completions/max_terminated_length": 990.0, "completions/mean_length": 449.0625, "completions/mean_terminated_length": 449.0625, "completions/min_length": 189.0, "completions/min_terminated_length": 189.0, "epoch": 1.5912, "grad_norm": 0.05441359058022499, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49073373.0, "reward": 0.7165328860282898, "reward_std": 0.2063169777393341, "rewards/accuracy_reward": 0.03125, "rewards/brier_reward": 0.4018140435218811, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7167187929153442, "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": 889.0, "completions/max_terminated_length": 889.0, "completions/mean_length": 422.125, "completions/mean_terminated_length": 422.125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.592, "grad_norm": 0.08532372117042542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49117381.0, "reward": 1.1334993839263916, "reward_std": 0.4015750288963318, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6888718605041504, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8837499618530273, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 835.0, "completions/mean_length": 494.359375, "completions/mean_terminated_length": 468.31146240234375, "completions/min_length": 299.0, "completions/min_terminated_length": 299.0, "epoch": 1.5928, "grad_norm": 0.07660417258739471, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49166492.0, "reward": 0.8544344902038574, "reward_std": 0.473730206489563, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4119921922683716, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8973437547683716, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 501.84375, "completions/mean_terminated_length": 476.1639099121094, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.5936, "grad_norm": 0.0824313685297966, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49214754.0, "reward": 0.9792648553848267, "reward_std": 0.3908240497112274, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5835281610488892, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8193749785423279, "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": 919.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 481.015625, "completions/mean_terminated_length": 481.015625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 1.5944, "grad_norm": 0.07570914924144745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49261715.0, "reward": 1.096282958984375, "reward_std": 0.4268569350242615, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.630064070224762, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9382812976837158, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 788.0, "completions/mean_length": 445.8125, "completions/mean_terminated_length": 417.37701416015625, "completions/min_length": 222.0, "completions/min_terminated_length": 222.0, "epoch": 1.5952, "grad_norm": 0.08457886427640915, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49306263.0, "reward": 1.0035220384597778, "reward_std": 0.33430972695350647, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5539171695709229, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8989062309265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 591.046875, "completions/mean_terminated_length": 577.0806274414062, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 1.596, "grad_norm": 0.037421390414237976, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49360122.0, "reward": 1.0682001113891602, "reward_std": 0.26157134771347046, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.683273434638977, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7326562404632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 783.0, "completions/mean_length": 402.125, "completions/mean_terminated_length": 392.2539978027344, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 1.5968, "grad_norm": 0.052837684750556946, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49400738.0, "reward": 1.0660446882247925, "reward_std": 0.25395679473876953, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6633374691009521, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8309375047683716, "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": 0.0, "completions/max_length": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 385.40625, "completions/mean_terminated_length": 385.40625, "completions/min_length": 240.0, "completions/min_terminated_length": 240.0, "epoch": 1.5976, "grad_norm": 0.07737142592668533, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49440876.0, "reward": 1.2057721614837646, "reward_std": 0.3051066994667053, "rewards/accuracy_reward": 0.671875, "rewards/brier_reward": 0.7396672368049622, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9298437833786011, "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": 1000.0, "completions/max_terminated_length": 1000.0, "completions/mean_length": 574.3125, "completions/mean_terminated_length": 574.3125, "completions/min_length": 287.0, "completions/min_terminated_length": 287.0, "epoch": 1.5984, "grad_norm": 0.05822407826781273, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49494256.0, "reward": 1.1232032775878906, "reward_std": 0.3625492453575134, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6526546478271484, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8982812166213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 555.8125, "completions/mean_terminated_length": 540.7096557617188, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 1.5992, "grad_norm": 0.055116280913352966, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49545636.0, "reward": 1.047455072402954, "reward_std": 0.38898319005966187, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.626157820224762, "rewards/confidence_one_or_zero": 0.34375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8460937142372131, "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": 0.0, "completions/max_length": 784.0, "completions/max_terminated_length": 784.0, "completions/mean_length": 367.140625, "completions/mean_terminated_length": 367.140625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.6, "grad_norm": 0.07245881110429764, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49585837.0, "reward": 0.968285322189331, "reward_std": 0.4345633089542389, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5459437370300293, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8700000047683716, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 582.515625, "completions/mean_terminated_length": 560.80322265625, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 1.6008, "grad_norm": 0.09007244557142258, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49639582.0, "reward": 0.7201400995254517, "reward_std": 0.2697876989841461, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.28402814269065857, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9025000333786011, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 981.0, "completions/mean_length": 566.015625, "completions/mean_terminated_length": 543.4917602539062, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "epoch": 1.6016, "grad_norm": 0.04329492524266243, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49693919.0, "reward": 0.9269930124282837, "reward_std": 0.3497583270072937, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.5727343559265137, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7484375238418579, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 748.0, "completions/mean_length": 504.109375, "completions/mean_terminated_length": 495.857177734375, "completions/min_length": 289.0, "completions/min_terminated_length": 289.0, "epoch": 1.6024, "grad_norm": 0.03762514516711235, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49741046.0, "reward": 0.9687579274177551, "reward_std": 0.1914844661951065, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.515639066696167, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9010937213897705, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 632.765625, "completions/mean_terminated_length": 620.1451416015625, "completions/min_length": 305.0, "completions/min_terminated_length": 305.0, "epoch": 1.6032, "grad_norm": 0.039013851433992386, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49798775.0, "reward": 0.7778069972991943, "reward_std": 0.32657599449157715, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.4306125044822693, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7478125095367432, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1009.0, "completions/mean_length": 615.40625, "completions/mean_terminated_length": 608.920654296875, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 1.604, "grad_norm": 0.04240768402814865, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49856401.0, "reward": 0.924507200717926, "reward_std": 0.36807483434677124, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.5365124940872192, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8078124523162842, "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": 730.0, "completions/max_terminated_length": 730.0, "completions/mean_length": 436.578125, "completions/mean_terminated_length": 436.578125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.6048, "grad_norm": 0.07216804474592209, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49900102.0, "reward": 0.7424236536026001, "reward_std": 0.38295140862464905, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3129703104496002, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8889062404632568, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 431.21875, "completions/mean_terminated_length": 421.8095397949219, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 1.6056, "grad_norm": 0.07923444360494614, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 49943604.0, "reward": 1.0461112260818481, "reward_std": 0.41375574469566345, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5922203063964844, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9110937118530273, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 502.921875, "completions/mean_terminated_length": 494.65081787109375, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 1.6064, "grad_norm": 0.05998127534985542, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 49991087.0, "reward": 0.9352422952651978, "reward_std": 0.374218225479126, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.49548280239105225, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8879687786102295, "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": 893.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 522.765625, "completions/mean_terminated_length": 522.765625, "completions/min_length": 349.0, "completions/min_terminated_length": 349.0, "epoch": 1.6072, "grad_norm": 0.040475718677043915, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50041120.0, "reward": 0.9074033498764038, "reward_std": 0.33642837405204773, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.48667967319488525, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8770312666893005, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 988.0, "completions/mean_length": 681.875, "completions/mean_terminated_length": 665.0491333007812, "completions/min_length": 352.0, "completions/min_terminated_length": 352.0, "epoch": 1.608, "grad_norm": 0.03730985149741173, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50101512.0, "reward": 0.7636877298355103, "reward_std": 0.2980477809906006, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.37112343311309814, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8379687666893005, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 942.0, "completions/mean_length": 433.375, "completions/mean_terminated_length": 394.0000305175781, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 1.6088, "grad_norm": 0.07365818321704865, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50144544.0, "reward": 0.9873057007789612, "reward_std": 0.30006611347198486, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.537109375, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8999999761581421, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 502.4375, "completions/mean_terminated_length": 485.6128845214844, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "epoch": 1.6096, "grad_norm": 0.04803185537457466, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50192252.0, "reward": 0.9273377060890198, "reward_std": 0.2921859323978424, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.49529844522476196, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8785938024520874, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 486.15625, "completions/mean_terminated_length": 477.61907958984375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.6104, "grad_norm": 0.06627459079027176, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50239558.0, "reward": 0.9605893492698669, "reward_std": 0.3887062072753906, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.499301552772522, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9251562356948853, "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": 794.0, "completions/max_terminated_length": 794.0, "completions/mean_length": 475.015625, "completions/mean_terminated_length": 475.015625, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.6112, "grad_norm": 0.031066402792930603, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50285063.0, "reward": 0.7591549754142761, "reward_std": 0.23470976948738098, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.34643280506134033, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8685938119888306, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1004.0, "completions/mean_length": 520.640625, "completions/mean_terminated_length": 512.6508178710938, "completions/min_length": 97.0, "completions/min_terminated_length": 97.0, "epoch": 1.612, "grad_norm": 0.13927718997001648, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50334800.0, "reward": 1.1036088466644287, "reward_std": 0.44266945123672485, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6603406071662903, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8918750286102295, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 425.296875, "completions/mean_terminated_length": 395.8524475097656, "completions/min_length": 194.0, "completions/min_terminated_length": 194.0, "epoch": 1.6128, "grad_norm": 0.07622174918651581, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50378595.0, "reward": 0.9078282117843628, "reward_std": 0.3752971589565277, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5656546950340271, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7423437237739563, "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": 0.0, "completions/max_length": 749.0, "completions/max_terminated_length": 749.0, "completions/mean_length": 341.6875, "completions/mean_terminated_length": 341.6875, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 1.6136, "grad_norm": 0.057932298630476, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50416863.0, "reward": 1.0825104713439941, "reward_std": 0.2648022174835205, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6025187373161316, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9490625262260437, "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": 918.0, "completions/max_terminated_length": 918.0, "completions/mean_length": 549.859375, "completions/mean_terminated_length": 549.859375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.6143999999999998, "grad_norm": 0.03972384333610535, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50468006.0, "reward": 0.8203510046005249, "reward_std": 0.30365216732025146, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.37507498264312744, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8912500143051147, "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": 865.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 497.875, "completions/mean_terminated_length": 497.875, "completions/min_length": 276.0, "completions/min_terminated_length": 276.0, "epoch": 1.6152, "grad_norm": 0.061932507902383804, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50519918.0, "reward": 0.9187354445457458, "reward_std": 0.4100819230079651, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4468437433242798, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9559375047683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 859.0, "completions/mean_length": 432.4375, "completions/mean_terminated_length": 413.3548278808594, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.616, "grad_norm": 0.07446826249361038, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50563818.0, "reward": 0.7908947467803955, "reward_std": 0.4037947654724121, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.37866249680519104, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8618749976158142, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 933.0, "completions/mean_length": 497.78125, "completions/mean_terminated_length": 471.901611328125, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 1.6168, "grad_norm": 0.07171833515167236, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50612236.0, "reward": 0.8222634792327881, "reward_std": 0.39653292298316956, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3788999915122986, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8984375, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 958.0, "completions/mean_length": 530.703125, "completions/mean_terminated_length": 522.873046875, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.6176, "grad_norm": 0.059156883507966995, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50661769.0, "reward": 0.9668237566947937, "reward_std": 0.37327682971954346, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.4961453080177307, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9470312595367432, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 592.015625, "completions/mean_terminated_length": 578.0806274414062, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 1.6183999999999998, "grad_norm": 0.05531245470046997, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50717098.0, "reward": 0.7634220123291016, "reward_std": 0.41741544008255005, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.3393421769142151, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8729687333106995, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 655.828125, "completions/mean_terminated_length": 637.7212524414062, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 1.6192, "grad_norm": 0.0408172570168972, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50775199.0, "reward": 0.7166118621826172, "reward_std": 0.2505551874637604, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.29259687662124634, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8887500166893005, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 559.40625, "completions/mean_terminated_length": 536.557373046875, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 1.62, "grad_norm": 0.04657100513577461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 50830249.0, "reward": 0.8206875324249268, "reward_std": 0.4113908112049103, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4382484555244446, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8076562881469727, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1021.0, "completions/mean_length": 601.890625, "completions/mean_terminated_length": 581.131103515625, "completions/min_length": 309.0, "completions/min_terminated_length": 309.0, "epoch": 1.6208, "grad_norm": 0.0690489411354065, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50885730.0, "reward": 0.8301079273223877, "reward_std": 0.43639564514160156, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.45708906650543213, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7887499928474426, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 707.953125, "completions/mean_terminated_length": 642.3585205078125, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.6216, "grad_norm": 0.042690377682447433, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50946703.0, "reward": 0.7389945983886719, "reward_std": 0.45572197437286377, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.35298749804496765, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.7887499928474426, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 908.0, "completions/mean_length": 517.203125, "completions/mean_terminated_length": 483.41668701171875, "completions/min_length": 227.0, "completions/min_terminated_length": 227.0, "epoch": 1.6223999999999998, "grad_norm": 0.07198864221572876, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 50995740.0, "reward": 0.89693683385849, "reward_std": 0.4755600392818451, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4657468795776367, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8598437309265137, "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": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 492.859375, "completions/mean_terminated_length": 492.859375, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.6232, "grad_norm": 0.04903532937169075, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51043763.0, "reward": 0.8047407865524292, "reward_std": 0.354175865650177, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.4063546657562256, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8514062762260437, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 570.0, "completions/mean_length": 376.1875, "completions/mean_terminated_length": 365.90478515625, "completions/min_length": 143.0, "completions/min_terminated_length": 143.0, "epoch": 1.624, "grad_norm": 0.13053588569164276, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51082767.0, "reward": 0.7783298492431641, "reward_std": 0.36787283420562744, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3535327911376953, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.897656261920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 740.0, "completions/mean_length": 403.875, "completions/mean_terminated_length": 394.0317687988281, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.6248, "grad_norm": 0.046062637120485306, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51124071.0, "reward": 1.0104886293411255, "reward_std": 0.22763967514038086, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5522249937057495, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9243749976158142, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 917.0, "completions/mean_length": 584.296875, "completions/mean_terminated_length": 562.672119140625, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.6256, "grad_norm": 0.03753700852394104, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51178458.0, "reward": 0.8575280904769897, "reward_std": 0.3245254456996918, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.5744296908378601, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.6142187714576721, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1001.0, "completions/mean_length": 548.515625, "completions/mean_terminated_length": 508.2203369140625, "completions/min_length": 130.0, "completions/min_terminated_length": 130.0, "epoch": 1.6263999999999998, "grad_norm": 0.050572704523801804, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51231371.0, "reward": 0.9510289430618286, "reward_std": 0.31915098428726196, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5583062171936035, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7931249737739563, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 636.6875, "completions/mean_terminated_length": 589.122802734375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.6272, "grad_norm": 0.04775494337081909, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51288263.0, "reward": 0.9483305811882019, "reward_std": 0.4547721743583679, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5529093742370605, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8031249642372131, "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": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 512.84375, "completions/mean_terminated_length": 512.84375, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 1.6280000000000001, "grad_norm": 0.05309322848916054, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51338013.0, "reward": 0.928358793258667, "reward_std": 0.37302446365356445, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4817156493663788, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8965624570846558, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 582.421875, "completions/mean_terminated_length": 575.4127197265625, "completions/min_length": 322.0, "completions/min_terminated_length": 322.0, "epoch": 1.6288, "grad_norm": 0.025434231385588646, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51391800.0, "reward": 0.7830438017845154, "reward_std": 0.2275787591934204, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.40983593463897705, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8057812452316284, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 542.734375, "completions/mean_terminated_length": 510.6500244140625, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 1.6296, "grad_norm": 0.04639815539121628, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51443143.0, "reward": 0.6374228000640869, "reward_std": 0.3367694020271301, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.2279687374830246, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.859375, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 531.234375, "completions/mean_terminated_length": 498.38336181640625, "completions/min_length": 302.0, "completions/min_terminated_length": 302.0, "epoch": 1.6303999999999998, "grad_norm": 0.06838368624448776, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51492950.0, "reward": 0.7476478815078735, "reward_std": 0.4558124542236328, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.3546687662601471, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8037500381469727, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 943.0, "completions/mean_length": 611.703125, "completions/mean_terminated_length": 569.0516967773438, "completions/min_length": 293.0, "completions/min_terminated_length": 293.0, "epoch": 1.6312, "grad_norm": 0.03917788341641426, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51548707.0, "reward": 0.995338499546051, "reward_std": 0.3531251847743988, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5687999725341797, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8515625, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 540.328125, "completions/mean_terminated_length": 499.3389892578125, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.6320000000000001, "grad_norm": 0.07080391049385071, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51599336.0, "reward": 0.9312205910682678, "reward_std": 0.43637222051620483, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5186890363693237, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8432812690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 446.15625, "completions/mean_terminated_length": 427.51611328125, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.6328, "grad_norm": 0.07214830070734024, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51645570.0, "reward": 0.9976939558982849, "reward_std": 0.32459065318107605, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.6360108852386475, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7295312881469727, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 998.0, "completions/mean_length": 512.59375, "completions/mean_terminated_length": 487.4425964355469, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.6336, "grad_norm": 0.049780212342739105, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51695368.0, "reward": 0.8019049763679504, "reward_std": 0.3396756052970886, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.3694327771663666, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8923437595367432, "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": 895.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 393.1875, "completions/mean_terminated_length": 393.1875, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.6343999999999999, "grad_norm": 0.08375740051269531, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51736180.0, "reward": 0.9269659519195557, "reward_std": 0.43577051162719727, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.47892966866493225, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9120312929153442, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 977.0, "completions/mean_length": 555.890625, "completions/mean_terminated_length": 532.8688354492188, "completions/min_length": 290.0, "completions/min_terminated_length": 290.0, "epoch": 1.6352, "grad_norm": 0.05589503049850464, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51788269.0, "reward": 1.007197618484497, "reward_std": 0.38032257556915283, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.6862687468528748, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.6699999570846558, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 976.0, "completions/mean_length": 508.3125, "completions/mean_terminated_length": 491.6773986816406, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.6360000000000001, "grad_norm": 0.05599916726350784, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51837121.0, "reward": 0.9697258472442627, "reward_std": 0.3504003584384918, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5957000255584717, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7915624976158142, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 941.0, "completions/mean_length": 579.015625, "completions/mean_terminated_length": 541.3051147460938, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.6368, "grad_norm": 0.061221957206726074, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51890146.0, "reward": 0.7671032547950745, "reward_std": 0.4145089387893677, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3623296916484833, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.850781261920929, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 448.390625, "completions/mean_terminated_length": 420.0819396972656, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 1.6376, "grad_norm": 0.0895153060555458, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 51933435.0, "reward": 1.0707659721374512, "reward_std": 0.4599606990814209, "rewards/accuracy_reward": 0.5625, "rewards/brier_reward": 0.6259046792984009, "rewards/confidence_one_or_zero": 0.171875, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.88671875, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1010.0, "completions/mean_length": 535.4375, "completions/mean_terminated_length": 494.0339050292969, "completions/min_length": 212.0, "completions/min_terminated_length": 212.0, "epoch": 1.6383999999999999, "grad_norm": 0.08086369931697845, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 51983031.0, "reward": 0.9674524068832397, "reward_std": 0.41636988520622253, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.606778085231781, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.762499988079071, "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": 995.0, "completions/mean_length": 529.390625, "completions/mean_terminated_length": 496.41668701171875, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 1.6392, "grad_norm": 0.05951106920838356, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52032320.0, "reward": 0.7201986312866211, "reward_std": 0.4322050213813782, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.284145325422287, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8860937356948853, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 779.0, "completions/mean_length": 512.515625, "completions/mean_terminated_length": 469.16949462890625, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 1.6400000000000001, "grad_norm": 0.05541546270251274, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52081713.0, "reward": 0.7413297891616821, "reward_std": 0.39424604177474976, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.3420328199863434, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8314062356948853, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 978.0, "completions/mean_length": 574.046875, "completions/mean_terminated_length": 551.9180297851562, "completions/min_length": 298.0, "completions/min_terminated_length": 298.0, "epoch": 1.6408, "grad_norm": 0.031132688745856285, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52134644.0, "reward": 0.862634539604187, "reward_std": 0.31174764037132263, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4440171718597412, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8567187786102295, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 986.0, "completions/mean_length": 526.390625, "completions/mean_terminated_length": 510.33868408203125, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 1.6416, "grad_norm": 0.03454352542757988, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52185869.0, "reward": 0.861190915107727, "reward_std": 0.28083229064941406, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.4098796844482422, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.91796875, "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": 689.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 421.21875, "completions/mean_terminated_length": 421.21875, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.6423999999999999, "grad_norm": 0.05955008417367935, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52229611.0, "reward": 1.1371915340423584, "reward_std": 0.3149794042110443, "rewards/accuracy_reward": 0.59375, "rewards/brier_reward": 0.6806312203407288, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9140625, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 505.90625, "completions/mean_terminated_length": 442.28070068359375, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.6432, "grad_norm": 0.04286055639386177, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52278437.0, "reward": 1.009696125984192, "reward_std": 0.35911041498184204, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5975155830383301, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8274999856948853, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 683.0, "completions/mean_length": 415.65625, "completions/mean_terminated_length": 396.0322570800781, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.6440000000000001, "grad_norm": 0.07089867442846298, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52321327.0, "reward": 1.039825201034546, "reward_std": 0.3707178831100464, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5796483755111694, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9179375171661377, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 546.171875, "completions/mean_terminated_length": 522.672119140625, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.6448, "grad_norm": 0.0525909923017025, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52371930.0, "reward": 0.8286423087120056, "reward_std": 0.4424474835395813, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.42290782928466797, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8417187929153442, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1003.0, "completions/mean_length": 529.203125, "completions/mean_terminated_length": 521.3492431640625, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 1.6456, "grad_norm": 0.033911217004060745, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52422615.0, "reward": 0.7825390100479126, "reward_std": 0.2702847719192505, "rewards/accuracy_reward": 0.078125, "rewards/brier_reward": 0.5025765895843506, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.6264062523841858, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 992.0, "completions/mean_length": 626.796875, "completions/mean_terminated_length": 613.9838256835938, "completions/min_length": 320.0, "completions/min_terminated_length": 320.0, "epoch": 1.6463999999999999, "grad_norm": 0.06342291086912155, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52478058.0, "reward": 0.8113032579421997, "reward_std": 0.35078203678131104, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.40385469794273376, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8501562476158142, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 450.796875, "completions/mean_terminated_length": 441.69842529296875, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 1.6472, "grad_norm": 0.048285841941833496, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52522573.0, "reward": 1.2221025228500366, "reward_std": 0.3403938412666321, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.8035781383514404, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8506250381469727, "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": 0.0, "completions/max_length": 930.0, "completions/max_terminated_length": 930.0, "completions/mean_length": 397.625, "completions/mean_terminated_length": 397.625, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.6480000000000001, "grad_norm": 0.0499635748565197, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52564693.0, "reward": 1.1604602336883545, "reward_std": 0.27371376752853394, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6802937984466553, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9415624737739563, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 820.0, "completions/mean_length": 473.859375, "completions/mean_terminated_length": 446.8032531738281, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.6488, "grad_norm": 0.0371604859828949, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52612012.0, "reward": 0.9695456027984619, "reward_std": 0.3085736036300659, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.5172140598297119, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9048437476158142, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 449.0625, "completions/mean_terminated_length": 420.786865234375, "completions/min_length": 155.0, "completions/min_terminated_length": 155.0, "epoch": 1.6496, "grad_norm": 0.12711186707019806, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52655552.0, "reward": 0.7778321504592896, "reward_std": 0.39436155557632446, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.33691251277923584, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8990625143051147, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 927.0, "completions/mean_length": 517.109375, "completions/mean_terminated_length": 464.67242431640625, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.6503999999999999, "grad_norm": 0.051150865852832794, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52704967.0, "reward": 0.9215124845504761, "reward_std": 0.32790619134902954, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.530523419380188, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7970312833786011, "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": 686.0, "completions/max_terminated_length": 686.0, "completions/mean_length": 318.28125, "completions/mean_terminated_length": 318.28125, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 1.6512, "grad_norm": 0.06042608991265297, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52740713.0, "reward": 1.1744049787521362, "reward_std": 0.2895887494087219, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.7081828117370605, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9342187643051147, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 945.0, "completions/mean_length": 522.625, "completions/mean_terminated_length": 489.20001220703125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.6520000000000001, "grad_norm": 0.05291437357664108, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 52790417.0, "reward": 0.8086134195327759, "reward_std": 0.3716716766357422, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.3984749913215637, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8474999666213989, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 459.46875, "completions/mean_terminated_length": 421.8333435058594, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.6528, "grad_norm": 0.05201862379908562, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52835823.0, "reward": 1.1321485042572021, "reward_std": 0.32269811630249023, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6861702799797058, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8870312571525574, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 946.0, "completions/mean_length": 474.453125, "completions/mean_terminated_length": 456.7257995605469, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.6536, "grad_norm": 0.04013447463512421, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52882524.0, "reward": 0.9832922220230103, "reward_std": 0.2972576916217804, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.6072077751159668, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7995312213897705, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 470.21875, "completions/mean_terminated_length": 461.4285888671875, "completions/min_length": 146.0, "completions/min_terminated_length": 146.0, "epoch": 1.6543999999999999, "grad_norm": 0.043885815888643265, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52928042.0, "reward": 0.7044845223426819, "reward_std": 0.2899695634841919, "rewards/accuracy_reward": 0.125, "rewards/brier_reward": 0.29959219694137573, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8520312309265137, "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": 984.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 486.96875, "completions/mean_terminated_length": 486.96875, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.6552, "grad_norm": 0.060188889503479004, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 52975496.0, "reward": 1.2338197231292725, "reward_std": 0.31529590487480164, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.7645124793052673, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9006249904632568, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 905.0, "completions/mean_length": 430.921875, "completions/mean_terminated_length": 411.7903137207031, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.6560000000000001, "grad_norm": 0.06281214952468872, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53021203.0, "reward": 0.9076799154281616, "reward_std": 0.3511761724948883, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4559828042984009, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9226562976837158, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 526.640625, "completions/mean_terminated_length": 475.1896667480469, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.6568, "grad_norm": 0.03600626066327095, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53072140.0, "reward": 0.7894666790962219, "reward_std": 0.3085775673389435, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.37580621242523193, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8528125286102295, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 950.0, "completions/mean_length": 491.203125, "completions/mean_terminated_length": 482.7460632324219, "completions/min_length": 221.0, "completions/min_terminated_length": 221.0, "epoch": 1.6576, "grad_norm": 0.06268244236707687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53118873.0, "reward": 0.7914806604385376, "reward_std": 0.411460816860199, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.33295938372612, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9415625333786011, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 934.0, "completions/mean_length": 526.953125, "completions/mean_terminated_length": 502.5081787109375, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 1.6583999999999999, "grad_norm": 0.08730834722518921, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53170006.0, "reward": 0.8251101970672607, "reward_std": 0.39696165919303894, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.46271875500679016, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7599999904632568, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 409.875, "completions/mean_terminated_length": 379.672119140625, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.6592, "grad_norm": 0.06963465362787247, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53211614.0, "reward": 1.0278149843215942, "reward_std": 0.44762295484542847, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.5556281805038452, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8331249952316284, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 871.0, "completions/mean_length": 524.921875, "completions/mean_terminated_length": 482.6271057128906, "completions/min_length": 260.0, "completions/min_terminated_length": 260.0, "epoch": 1.6600000000000001, "grad_norm": 0.04602630436420441, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53262169.0, "reward": 0.814422607421875, "reward_std": 0.3441695272922516, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4569687247276306, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.744687557220459, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 545.328125, "completions/mean_terminated_length": 529.8870849609375, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 1.6608, "grad_norm": 0.05605623126029968, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53313102.0, "reward": 0.9178587198257446, "reward_std": 0.3726399838924408, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5075905919075012, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8690625429153442, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 991.0, "completions/mean_length": 504.109375, "completions/mean_terminated_length": 495.857177734375, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 1.6616, "grad_norm": 0.05597860738635063, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53361621.0, "reward": 1.0472025871276855, "reward_std": 0.2987854480743408, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6725281476974487, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.776562511920929, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 611.484375, "completions/mean_terminated_length": 598.1774291992188, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 1.6623999999999999, "grad_norm": 0.03259315714240074, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53417252.0, "reward": 0.7427001595497131, "reward_std": 0.26837700605392456, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.34477347135543823, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8574999570846558, "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": 0.0, "completions/max_length": 962.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 474.546875, "completions/mean_terminated_length": 474.546875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.6632, "grad_norm": 0.055504146963357925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53463015.0, "reward": 0.9234939217567444, "reward_std": 0.3551004230976105, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.518860936164856, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8595312237739563, "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": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 437.96875, "completions/mean_terminated_length": 437.96875, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "epoch": 1.6640000000000001, "grad_norm": 0.06312746554613113, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53506229.0, "reward": 0.9086111783981323, "reward_std": 0.35981225967407227, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4422203004360199, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.959531307220459, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 919.0, "completions/mean_length": 527.671875, "completions/mean_terminated_length": 466.71929931640625, "completions/min_length": 231.0, "completions/min_terminated_length": 231.0, "epoch": 1.6648, "grad_norm": 0.06281504780054092, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53557328.0, "reward": 0.9997532367706299, "reward_std": 0.4671339988708496, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6088796854019165, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7873437404632568, "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": 867.0, "completions/max_terminated_length": 867.0, "completions/mean_length": 450.921875, "completions/mean_terminated_length": 450.921875, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 1.6656, "grad_norm": 0.07797171175479889, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53601483.0, "reward": 1.1620368957519531, "reward_std": 0.24964189529418945, "rewards/accuracy_reward": 0.640625, "rewards/brier_reward": 0.6990718841552734, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9404687285423279, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 757.0, "completions/mean_length": 460.09375, "completions/mean_terminated_length": 432.3606262207031, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.6663999999999999, "grad_norm": 0.057444531470537186, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 53647505.0, "reward": 0.9164687395095825, "reward_std": 0.39462482929229736, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.5516859292984009, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7214062213897705, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 987.0, "completions/mean_length": 526.265625, "completions/mean_terminated_length": 518.3651123046875, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 1.6672, "grad_norm": 0.06437806785106659, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53696722.0, "reward": 0.8689016699790955, "reward_std": 0.3944408595561981, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.44092655181884766, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8882812261581421, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 893.0, "completions/mean_length": 476.1875, "completions/mean_terminated_length": 458.51611328125, "completions/min_length": 242.0, "completions/min_terminated_length": 242.0, "epoch": 1.6680000000000001, "grad_norm": 0.06166909635066986, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53742878.0, "reward": 0.8780908584594727, "reward_std": 0.4374507665634155, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.45930469036102295, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8645312786102295, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 926.0, "completions/mean_length": 525.796875, "completions/mean_terminated_length": 509.7257995605469, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.6688, "grad_norm": 0.042499374598264694, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53794593.0, "reward": 1.056769847869873, "reward_std": 0.3193669021129608, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.6135375499725342, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.9017187356948853, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 427.75, "completions/mean_terminated_length": 408.51611328125, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.6696, "grad_norm": 0.06510841846466064, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53837313.0, "reward": 1.139875888824463, "reward_std": 0.35709500312805176, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.7016249895095825, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8759374618530273, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 864.0, "completions/mean_length": 531.03125, "completions/mean_terminated_length": 489.2542419433594, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.6703999999999999, "grad_norm": 0.0573340468108654, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53888451.0, "reward": 0.6475964784622192, "reward_std": 0.2605069577693939, "rewards/accuracy_reward": 0.140625, "rewards/brier_reward": 0.23269063234329224, "rewards/confidence_one_or_zero": 0.28125, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8534374833106995, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 879.0, "completions/mean_length": 524.046875, "completions/mean_terminated_length": 507.9193420410156, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 1.6712, "grad_norm": 0.050325069576501846, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53938438.0, "reward": 0.837993860244751, "reward_std": 0.4010744094848633, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.4416109323501587, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8370312452316284, "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": 0.0, "completions/max_length": 716.0, "completions/max_terminated_length": 716.0, "completions/mean_length": 426.296875, "completions/mean_terminated_length": 426.296875, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "epoch": 1.6720000000000002, "grad_norm": 0.03519872948527336, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 53985449.0, "reward": 0.7444063425064087, "reward_std": 0.1577373892068863, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.3325609266757965, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8857812881469727, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 486.6875, "completions/mean_terminated_length": 450.86669921875, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.6728, "grad_norm": 0.04092448204755783, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54032565.0, "reward": 0.7287142276763916, "reward_std": 0.3244425058364868, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.3168015480041504, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8501562476158142, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 608.0, "completions/mean_length": 343.25, "completions/mean_terminated_length": 332.4444580078125, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 1.6736, "grad_norm": 0.07899388670921326, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54071461.0, "reward": 1.1614723205566406, "reward_std": 0.4717252552509308, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.6823171973228455, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9339062571525574, "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": 0.0, "completions/max_length": 925.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 509.921875, "completions/mean_terminated_length": 509.921875, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 1.6743999999999999, "grad_norm": 0.03727780282497406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54121008.0, "reward": 0.9708839058876038, "reward_std": 0.3103753924369812, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5198906660079956, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9340624809265137, "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": 758.0, "completions/max_terminated_length": 758.0, "completions/mean_length": 458.734375, "completions/mean_terminated_length": 458.734375, "completions/min_length": 267.0, "completions/min_terminated_length": 267.0, "epoch": 1.6752, "grad_norm": 0.042891260236501694, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54168671.0, "reward": 0.9543110728263855, "reward_std": 0.23186130821704865, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.596120297908783, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7610937356948853, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 928.0, "completions/mean_length": 464.03125, "completions/mean_terminated_length": 455.14288330078125, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 1.6760000000000002, "grad_norm": 0.058262355625629425, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54215985.0, "reward": 0.7704477310180664, "reward_std": 0.3692588210105896, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.3846437335014343, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8334375023841858, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 955.0, "completions/mean_length": 560.375, "completions/mean_terminated_length": 512.413818359375, "completions/min_length": 303.0, "completions/min_terminated_length": 303.0, "epoch": 1.6768, "grad_norm": 0.056088365614414215, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54268633.0, "reward": 0.9062469601631165, "reward_std": 0.43441832065582275, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4999921917915344, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8245312571525574, "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": 823.0, "completions/max_terminated_length": 823.0, "completions/mean_length": 364.28125, "completions/mean_terminated_length": 364.28125, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.6776, "grad_norm": 0.09470509737730026, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54308571.0, "reward": 1.3329994678497314, "reward_std": 0.3562640845775604, "rewards/accuracy_reward": 0.828125, "rewards/brier_reward": 0.8378719091415405, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9625000357627869, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 577.328125, "completions/mean_terminated_length": 547.550048828125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 1.6784, "grad_norm": 0.04592433199286461, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54362224.0, "reward": 0.7530375719070435, "reward_std": 0.35326915979385376, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.3810734450817108, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8020312786102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1015.0, "completions/mean_length": 475.625, "completions/mean_terminated_length": 439.0666809082031, "completions/min_length": 182.0, "completions/min_terminated_length": 182.0, "epoch": 1.6792, "grad_norm": 0.07934398949146271, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54408840.0, "reward": 0.9078010320663452, "reward_std": 0.4656458795070648, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5187249779701233, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.815625011920929, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 921.0, "completions/mean_length": 539.796875, "completions/mean_terminated_length": 489.7069091796875, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.6800000000000002, "grad_norm": 0.0717938095331192, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54459243.0, "reward": 0.8197203874588013, "reward_std": 0.4730215072631836, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.42068904638290405, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.819531261920929, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 452.1875, "completions/mean_terminated_length": 433.7419128417969, "completions/min_length": 208.0, "completions/min_terminated_length": 208.0, "epoch": 1.6808, "grad_norm": 0.06923610717058182, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54504471.0, "reward": 1.0572211742401123, "reward_std": 0.3884880542755127, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.6613156199455261, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8028124570846558, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 892.0, "completions/mean_length": 478.796875, "completions/mean_terminated_length": 442.45001220703125, "completions/min_length": 181.0, "completions/min_terminated_length": 181.0, "epoch": 1.6816, "grad_norm": 0.05978574976325035, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54551226.0, "reward": 0.6824023723602295, "reward_std": 0.3384862542152405, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.31792810559272766, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7865625023841858, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 868.0, "completions/mean_length": 540.609375, "completions/mean_terminated_length": 525.01611328125, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 1.6824, "grad_norm": 0.06829936802387238, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54602625.0, "reward": 0.9372634291648865, "reward_std": 0.39005154371261597, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4995250105857849, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8871874809265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 494.21875, "completions/mean_terminated_length": 477.1290283203125, "completions/min_length": 273.0, "completions/min_terminated_length": 273.0, "epoch": 1.6832, "grad_norm": 0.06648819893598557, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54652159.0, "reward": 0.9490790367126465, "reward_std": 0.3058614730834961, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.5544062852859497, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8464062213897705, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 616.375, "completions/mean_terminated_length": 581.8305053710938, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 1.6840000000000002, "grad_norm": 0.02174488455057144, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54708055.0, "reward": 0.8914171457290649, "reward_std": 0.2386280596256256, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.5484578013420105, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7342187166213989, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 500.046875, "completions/mean_terminated_length": 445.8448181152344, "completions/min_length": 99.0, "completions/min_terminated_length": 99.0, "epoch": 1.6848, "grad_norm": 0.10488104075193405, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54757482.0, "reward": 1.0274267196655273, "reward_std": 0.4606102406978607, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6642265319824219, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.7568594217300415, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 852.0, "completions/mean_length": 390.375, "completions/mean_terminated_length": 380.3174743652344, "completions/min_length": 213.0, "completions/min_terminated_length": 213.0, "epoch": 1.6856, "grad_norm": 0.058493949472904205, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54800322.0, "reward": 1.1752712726593018, "reward_std": 0.24954929947853088, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7099156379699707, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9137499928474426, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 865.0, "completions/mean_length": 533.21875, "completions/mean_terminated_length": 517.3870849609375, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 1.6864, "grad_norm": 0.043576568365097046, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 54851072.0, "reward": 1.1775321960449219, "reward_std": 0.3108685612678528, "rewards/accuracy_reward": 0.671875, "rewards/brier_reward": 0.7144374847412109, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9193750023841858, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 663.75, "completions/mean_terminated_length": 619.5087890625, "completions/min_length": 286.0, "completions/min_terminated_length": 286.0, "epoch": 1.6872, "grad_norm": 0.03843590244650841, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54909040.0, "reward": 0.8387594223022461, "reward_std": 0.4233526587486267, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.45876720547676086, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.78515625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1018.0, "completions/mean_length": 618.8125, "completions/mean_terminated_length": 591.800048828125, "completions/min_length": 354.0, "completions/min_terminated_length": 354.0, "epoch": 1.688, "grad_norm": 0.03134158253669739, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 54965460.0, "reward": 0.6362438201904297, "reward_std": 0.28373926877975464, "rewards/accuracy_reward": 0.0625, "rewards/brier_reward": 0.2724859416484833, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7632812261581421, "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": 682.0, "completions/max_terminated_length": 682.0, "completions/mean_length": 360.828125, "completions/mean_terminated_length": 360.828125, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.6888, "grad_norm": 0.051607273519039154, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55006377.0, "reward": 1.2251713275909424, "reward_std": 0.3065217137336731, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7315906286239624, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.966546893119812, "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": 719.0, "completions/max_terminated_length": 719.0, "completions/mean_length": 394.890625, "completions/mean_terminated_length": 394.890625, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.6896, "grad_norm": 0.084925077855587, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55047938.0, "reward": 1.0577292442321777, "reward_std": 0.44016095995903015, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.5842062830924988, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9431250095367432, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 938.0, "completions/mean_length": 601.9375, "completions/mean_terminated_length": 588.3225708007812, "completions/min_length": 206.0, "completions/min_terminated_length": 206.0, "epoch": 1.6904, "grad_norm": 0.049864839762449265, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55102446.0, "reward": 0.7574814558029175, "reward_std": 0.29864540696144104, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.3118359446525574, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9164062738418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 995.0, "completions/mean_length": 477.34375, "completions/mean_terminated_length": 459.70965576171875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "epoch": 1.6912, "grad_norm": 0.04524515941739082, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55148980.0, "reward": 0.8945485949516296, "reward_std": 0.2824854850769043, "rewards/accuracy_reward": 0.359375, "rewards/brier_reward": 0.46097031235694885, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8814062476158142, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 603.71875, "completions/mean_terminated_length": 552.1052856445312, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 1.692, "grad_norm": 0.03301442041993141, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55205186.0, "reward": 0.7923517227172852, "reward_std": 0.23132118582725525, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.3815765678882599, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8490625023841858, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 947.0, "completions/mean_length": 571.828125, "completions/mean_terminated_length": 516.2982788085938, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 1.6928, "grad_norm": 0.07342565804719925, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55258935.0, "reward": 0.7545742988586426, "reward_std": 0.3559301495552063, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.35289686918258667, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.8378124833106995, "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": 0.125, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 474.21875, "completions/mean_terminated_length": 395.6785888671875, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.6936, "grad_norm": 0.06970228999853134, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55304485.0, "reward": 0.8822531700134277, "reward_std": 0.43868961930274963, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.46762967109680176, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.875, "rewards/mean_confidence_reward": 0.7845312356948853, "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": 0.0, "completions/max_length": 661.0, "completions/max_terminated_length": 661.0, "completions/mean_length": 439.296875, "completions/mean_terminated_length": 439.296875, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 1.6944, "grad_norm": 0.05405895411968231, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55347912.0, "reward": 1.2209010124206543, "reward_std": 0.41977912187576294, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.7386749982833862, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.953125, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 900.0, "completions/mean_length": 397.890625, "completions/mean_terminated_length": 377.69354248046875, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.6952, "grad_norm": 0.08239223808050156, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55390321.0, "reward": 0.9400032758712769, "reward_std": 0.4288216829299927, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.4893797039985657, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.9089062809944153, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 695.0, "completions/mean_length": 382.1875, "completions/mean_terminated_length": 372.0000305175781, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 1.696, "grad_norm": 0.07981947809457779, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55429901.0, "reward": 1.1171009540557861, "reward_std": 0.33978259563446045, "rewards/accuracy_reward": 0.609375, "rewards/brier_reward": 0.6404500007629395, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9440625309944153, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 973.0, "completions/mean_length": 554.40625, "completions/mean_terminated_length": 523.1000366210938, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.6968, "grad_norm": 0.05872069299221039, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55482311.0, "reward": 0.8912938833236694, "reward_std": 0.39807188510894775, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4700859487056732, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8489062786102295, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 497.09375, "completions/mean_terminated_length": 480.0967712402344, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.6976, "grad_norm": 0.06940463185310364, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55529629.0, "reward": 0.7875813245773315, "reward_std": 0.384212851524353, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.4189109206199646, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7859375476837158, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 895.0, "completions/mean_length": 530.390625, "completions/mean_terminated_length": 497.4833679199219, "completions/min_length": 209.0, "completions/min_terminated_length": 209.0, "epoch": 1.6984, "grad_norm": 0.08665399998426437, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55579286.0, "reward": 1.0720642805099487, "reward_std": 0.4732738733291626, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.6441265344619751, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8776562809944153, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 971.0, "completions/mean_length": 512.890625, "completions/mean_terminated_length": 496.4031982421875, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 1.6992, "grad_norm": 0.07498368620872498, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55628927.0, "reward": 0.9692164659500122, "reward_std": 0.43516719341278076, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.5321812629699707, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8796875476837158, "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": 0.0, "completions/max_length": 803.0, "completions/max_terminated_length": 803.0, "completions/mean_length": 436.640625, "completions/mean_terminated_length": 436.640625, "completions/min_length": 217.0, "completions/min_terminated_length": 217.0, "epoch": 1.7, "grad_norm": 0.07539547234773636, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55674712.0, "reward": 0.9882837533950806, "reward_std": 0.4196791648864746, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.5234405994415283, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9378124475479126, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 863.0, "completions/mean_length": 424.875, "completions/mean_terminated_length": 415.3651123046875, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.7008, "grad_norm": 0.04790883511304855, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55717296.0, "reward": 0.7216488122940063, "reward_std": 0.25320881605148315, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.28704530000686646, "rewards/confidence_one_or_zero": 0.3125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.897656261920929, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 802.0, "completions/mean_length": 487.96875, "completions/mean_terminated_length": 479.4603576660156, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 1.7016, "grad_norm": 0.03692561760544777, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55764734.0, "reward": 0.8702665567398071, "reward_std": 0.28244245052337646, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4436562657356262, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8662500381469727, "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": 0.15625, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 605.953125, "completions/mean_terminated_length": 528.5370483398438, "completions/min_length": 225.0, "completions/min_terminated_length": 225.0, "epoch": 1.7024, "grad_norm": 0.0631670355796814, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55820715.0, "reward": 0.8798556327819824, "reward_std": 0.3666148781776428, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.49408435821533203, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.84375, "rewards/mean_confidence_reward": 0.7809375524520874, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1020.0, "completions/mean_length": 584.703125, "completions/mean_terminated_length": 555.4166870117188, "completions/min_length": 263.0, "completions/min_terminated_length": 263.0, "epoch": 1.7032, "grad_norm": 0.03651810437440872, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55875016.0, "reward": 0.763336718082428, "reward_std": 0.3016820549964905, "rewards/accuracy_reward": 0.1875, "rewards/brier_reward": 0.4016718566417694, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.7378125190734863, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 648.234375, "completions/mean_terminated_length": 570.2453002929688, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 1.704, "grad_norm": 0.04725270718336105, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 55932695.0, "reward": 0.8638032078742981, "reward_std": 0.42870500683784485, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.5088546872138977, "rewards/confidence_one_or_zero": 0.203125, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.7123437523841858, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 855.0, "completions/mean_length": 393.703125, "completions/mean_terminated_length": 383.69842529296875, "completions/min_length": 125.0, "completions/min_terminated_length": 125.0, "epoch": 1.7048, "grad_norm": 0.0581551119685173, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 55974692.0, "reward": 0.9405965805053711, "reward_std": 0.24285608530044556, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.4593156576156616, "rewards/confidence_one_or_zero": 0.296875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9615625143051147, "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": 854.0, "completions/max_terminated_length": 854.0, "completions/mean_length": 434.5625, "completions/mean_terminated_length": 434.5625, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.7056, "grad_norm": 0.04544590041041374, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56017816.0, "reward": 1.2561993598937988, "reward_std": 0.28912216424942017, "rewards/accuracy_reward": 0.71875, "rewards/brier_reward": 0.7936468720436096, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9203125238418579, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 957.0, "completions/mean_length": 496.6875, "completions/mean_terminated_length": 479.6773986816406, "completions/min_length": 261.0, "completions/min_terminated_length": 261.0, "epoch": 1.7064, "grad_norm": 0.04299485310912132, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56065220.0, "reward": 0.6536563634872437, "reward_std": 0.21125459671020508, "rewards/accuracy_reward": 0.109375, "rewards/brier_reward": 0.22918593883514404, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8904687166213989, "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": 902.0, "completions/max_terminated_length": 902.0, "completions/mean_length": 492.09375, "completions/mean_terminated_length": 492.09375, "completions/min_length": 232.0, "completions/min_terminated_length": 232.0, "epoch": 1.7072, "grad_norm": 0.09451844543218613, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56114218.0, "reward": 0.8497813940048218, "reward_std": 0.30817124247550964, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.5120609402656555, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7267187833786011, "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": 870.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 519.96875, "completions/mean_terminated_length": 519.96875, "completions/min_length": 258.0, "completions/min_terminated_length": 258.0, "epoch": 1.708, "grad_norm": 0.050140630453825, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56163704.0, "reward": 0.8371391892433167, "reward_std": 0.30311712622642517, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.40865159034729004, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8926562666893005, "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": 0.140625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1002.0, "completions/mean_length": 585.984375, "completions/mean_terminated_length": 514.30908203125, "completions/min_length": 201.0, "completions/min_terminated_length": 201.0, "epoch": 1.7088, "grad_norm": 0.07802070677280426, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56218359.0, "reward": 0.702332079410553, "reward_std": 0.34261009097099304, "rewards/accuracy_reward": 0.234375, "rewards/brier_reward": 0.31091248989105225, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.7921874523162842, "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": 829.0, "completions/max_terminated_length": 829.0, "completions/mean_length": 396.78125, "completions/mean_terminated_length": 396.78125, "completions/min_length": 229.0, "completions/min_terminated_length": 229.0, "epoch": 1.7096, "grad_norm": 0.06293534487485886, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56259913.0, "reward": 0.9815133810043335, "reward_std": 0.331825315952301, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5411499738693237, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8968750238418579, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 999.0, "completions/mean_length": 708.640625, "completions/mean_terminated_length": 643.1886596679688, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.7104, "grad_norm": 0.06238844245672226, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56322562.0, "reward": 0.6673756837844849, "reward_std": 0.34377002716064453, "rewards/accuracy_reward": 0.09375, "rewards/brier_reward": 0.4128749966621399, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.5709375143051147, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 870.0, "completions/mean_length": 444.765625, "completions/mean_terminated_length": 426.08062744140625, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.7112, "grad_norm": 0.06636221706867218, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56367155.0, "reward": 1.0204792022705078, "reward_std": 0.3834249973297119, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.6034562587738037, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8246874809265137, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 466.3125, "completions/mean_terminated_length": 438.8852233886719, "completions/min_length": 251.0, "completions/min_terminated_length": 251.0, "epoch": 1.712, "grad_norm": 0.06511831283569336, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56412871.0, "reward": 0.9884477257728577, "reward_std": 0.3660928010940552, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5393937826156616, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.8778125047683716, "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": 0.171875, "completions/max_length": 1024.0, "completions/max_terminated_length": 994.0, "completions/mean_length": 597.671875, "completions/mean_terminated_length": 509.1886901855469, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.7128, "grad_norm": 0.09538278728723526, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56467906.0, "reward": 0.7086725234985352, "reward_std": 0.4500133693218231, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.38609376549720764, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.828125, "rewards/mean_confidence_reward": 0.6234374642372131, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 883.0, "completions/mean_length": 489.640625, "completions/mean_terminated_length": 481.15875244140625, "completions/min_length": 288.0, "completions/min_terminated_length": 288.0, "epoch": 1.7136, "grad_norm": 0.050418704748153687, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56515707.0, "reward": 0.8993477821350098, "reward_std": 0.30176663398742676, "rewards/accuracy_reward": 0.328125, "rewards/brier_reward": 0.4861937165260315, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8428125381469727, "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": 0.0, "completions/max_length": 937.0, "completions/max_terminated_length": 937.0, "completions/mean_length": 499.0, "completions/mean_terminated_length": 499.0, "completions/min_length": 333.0, "completions/min_terminated_length": 333.0, "epoch": 1.7144, "grad_norm": 0.05980425700545311, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56564571.0, "reward": 1.1005642414093018, "reward_std": 0.37162572145462036, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6698765158653259, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.879812479019165, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 984.0, "completions/mean_length": 506.0625, "completions/mean_terminated_length": 471.5333557128906, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.7151999999999998, "grad_norm": 0.07765575498342514, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56614175.0, "reward": 0.9888696670532227, "reward_std": 0.454745352268219, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.571487545967102, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8503124713897705, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 954.0, "completions/mean_length": 516.78125, "completions/mean_terminated_length": 473.7966003417969, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.716, "grad_norm": 0.06963998824357986, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56663297.0, "reward": 1.0347133874893188, "reward_std": 0.39667391777038574, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.6475499868392944, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.7918750047683716, "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": 906.0, "completions/max_terminated_length": 906.0, "completions/mean_length": 347.71875, "completions/mean_terminated_length": 347.71875, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "epoch": 1.7168, "grad_norm": 0.08182205259799957, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56700367.0, "reward": 0.9182401895523071, "reward_std": 0.39932748675346375, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4458531141281128, "rewards/confidence_one_or_zero": 0.234375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9606249928474426, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 831.0, "completions/mean_length": 458.703125, "completions/mean_terminated_length": 449.7301940917969, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 1.7176, "grad_norm": 0.053823888301849365, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56746236.0, "reward": 1.0216079950332642, "reward_std": 0.39521700143814087, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5744640827178955, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8692187666893005, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 983.0, "completions/mean_length": 557.796875, "completions/mean_terminated_length": 542.758056640625, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "epoch": 1.7184, "grad_norm": 0.03979848697781563, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56799583.0, "reward": 0.9847251176834106, "reward_std": 0.2570076584815979, "rewards/accuracy_reward": 0.453125, "rewards/brier_reward": 0.547573447227478, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8826562166213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1024.0, "completions/mean_length": 501.625, "completions/mean_terminated_length": 484.774169921875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 1.7191999999999998, "grad_norm": 0.09170826524496078, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56847175.0, "reward": 0.7629424929618835, "reward_std": 0.4466031491756439, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.3383828103542328, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.8560937643051147, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 689.0, "completions/mean_length": 457.890625, "completions/mean_terminated_length": 448.90478515625, "completions/min_length": 237.0, "completions/min_terminated_length": 237.0, "epoch": 1.72, "grad_norm": 0.049218542873859406, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 56895680.0, "reward": 0.7526338696479797, "reward_std": 0.35139238834381104, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.30214062333106995, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9234374761581421, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1006.0, "completions/mean_length": 516.59375, "completions/mean_terminated_length": 491.6393127441406, "completions/min_length": 228.0, "completions/min_terminated_length": 228.0, "epoch": 1.7208, "grad_norm": 0.0770399272441864, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56944790.0, "reward": 0.8322603106498718, "reward_std": 0.4375596344470978, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.4613937437534332, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7831250429153442, "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": 965.0, "completions/max_terminated_length": 965.0, "completions/mean_length": 374.15625, "completions/mean_terminated_length": 374.15625, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "epoch": 1.7216, "grad_norm": 0.13332317769527435, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 56983808.0, "reward": 1.2307915687561035, "reward_std": 0.24547775089740753, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.8053311109542847, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8667656183242798, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 982.0, "completions/mean_length": 568.703125, "completions/mean_terminated_length": 561.4761962890625, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 1.7224, "grad_norm": 0.047779280692338943, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57038189.0, "reward": 1.0120360851287842, "reward_std": 0.4083789587020874, "rewards/accuracy_reward": 0.4375, "rewards/brier_reward": 0.6021953225135803, "rewards/confidence_one_or_zero": 0.078125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8379687070846558, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 925.0, "completions/mean_length": 416.3125, "completions/mean_terminated_length": 406.66668701171875, "completions/min_length": 190.0, "completions/min_terminated_length": 190.0, "epoch": 1.7231999999999998, "grad_norm": 0.05109807476401329, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57080801.0, "reward": 1.0413986444473267, "reward_std": 0.3169917166233063, "rewards/accuracy_reward": 0.5, "rewards/brier_reward": 0.5984203219413757, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9032812118530273, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 996.0, "completions/mean_length": 543.71875, "completions/mean_terminated_length": 528.2257690429688, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 1.724, "grad_norm": 0.04312647506594658, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57133183.0, "reward": 0.7606228590011597, "reward_std": 0.3447888493537903, "rewards/accuracy_reward": 0.171875, "rewards/brier_reward": 0.38061872124671936, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7825000286102295, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 956.0, "completions/mean_length": 562.890625, "completions/mean_terminated_length": 532.1500244140625, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.7248, "grad_norm": 0.1052844300866127, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57189880.0, "reward": 0.8962485790252686, "reward_std": 0.5143172740936279, "rewards/accuracy_reward": 0.40625, "rewards/brier_reward": 0.4799953103065491, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8654687404632568, "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": 1013.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 418.765625, "completions/mean_terminated_length": 418.765625, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 1.7256, "grad_norm": 0.07555339485406876, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57232233.0, "reward": 0.8651635050773621, "reward_std": 0.3125537633895874, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.46469998359680176, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.8350000381469727, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 989.0, "completions/mean_length": 462.078125, "completions/mean_terminated_length": 453.15875244140625, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 1.7264, "grad_norm": 0.0613536573946476, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57279534.0, "reward": 0.9769971370697021, "reward_std": 0.33788132667541504, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5946171879768372, "rewards/confidence_one_or_zero": 0.25, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.7829687595367432, "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": 878.0, "completions/max_terminated_length": 878.0, "completions/mean_length": 446.984375, "completions/mean_terminated_length": 446.984375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 1.7271999999999998, "grad_norm": 0.03352764621376991, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57325245.0, "reward": 1.0986757278442383, "reward_std": 0.18391025066375732, "rewards/accuracy_reward": 0.46875, "rewards/brier_reward": 0.728600025177002, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.7568749785423279, "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": 1014.0, "completions/max_terminated_length": 1014.0, "completions/mean_length": 486.21875, "completions/mean_terminated_length": 486.21875, "completions/min_length": 176.0, "completions/min_terminated_length": 176.0, "epoch": 1.728, "grad_norm": 0.05196971073746681, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57373179.0, "reward": 0.7114181518554688, "reward_std": 0.3318321108818054, "rewards/accuracy_reward": 0.15625, "rewards/brier_reward": 0.2665843963623047, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9278124570846558, "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": 773.0, "completions/max_terminated_length": 773.0, "completions/mean_length": 480.921875, "completions/mean_terminated_length": 480.921875, "completions/min_length": 291.0, "completions/min_terminated_length": 291.0, "epoch": 1.7288000000000001, "grad_norm": 0.06884542852640152, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57423350.0, "reward": 0.9745498895645142, "reward_std": 0.2857052683830261, "rewards/accuracy_reward": 0.28125, "rewards/brier_reward": 0.6678484678268433, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.6614062786102295, "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": 874.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 362.734375, "completions/mean_terminated_length": 362.734375, "completions/min_length": 103.0, "completions/min_terminated_length": 103.0, "epoch": 1.7296, "grad_norm": 0.069858118891716, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57463205.0, "reward": 0.8303964138031006, "reward_std": 0.23489007353782654, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.3482903242111206, "rewards/confidence_one_or_zero": 0.21875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9763749837875366, "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": 747.0, "completions/max_terminated_length": 747.0, "completions/mean_length": 389.328125, "completions/mean_terminated_length": 389.328125, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "epoch": 1.7304, "grad_norm": 0.05479664355516434, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57504266.0, "reward": 1.2042438983917236, "reward_std": 0.2820567190647125, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7522359490394592, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9064062833786011, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 964.0, "completions/mean_length": 594.96875, "completions/mean_terminated_length": 550.586181640625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 1.7311999999999999, "grad_norm": 0.04756837338209152, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57558536.0, "reward": 0.8563711047172546, "reward_std": 0.3218035101890564, "rewards/accuracy_reward": 0.265625, "rewards/brier_reward": 0.5408656597137451, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.6753125190734863, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1016.0, "completions/mean_length": 569.53125, "completions/mean_terminated_length": 531.0169677734375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 1.732, "grad_norm": 0.06605034321546555, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57612106.0, "reward": 0.9602898359298706, "reward_std": 0.3812818229198456, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.623703122138977, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.921875, "rewards/mean_confidence_reward": 0.6893750429153442, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 448.015625, "completions/mean_terminated_length": 438.873046875, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 1.7328000000000001, "grad_norm": 0.06551092118024826, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57658395.0, "reward": 0.8079315423965454, "reward_std": 0.33709394931793213, "rewards/accuracy_reward": 0.25, "rewards/brier_reward": 0.3814859390258789, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8982812166213989, "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": 739.0, "completions/max_terminated_length": 739.0, "completions/mean_length": 408.125, "completions/mean_terminated_length": 408.125, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 1.7336, "grad_norm": 0.05839702859520912, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57700547.0, "reward": 1.0739150047302246, "reward_std": 0.3328242301940918, "rewards/accuracy_reward": 0.53125, "rewards/brier_reward": 0.6165781021118164, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9068750143051147, "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": 866.0, "completions/max_terminated_length": 866.0, "completions/mean_length": 423.234375, "completions/mean_terminated_length": 423.234375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 1.7344, "grad_norm": 0.09010688960552216, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57744706.0, "reward": 1.1140345335006714, "reward_std": 0.2084733545780182, "rewards/accuracy_reward": 0.578125, "rewards/brier_reward": 0.64994215965271, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9279687404632568, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 1019.0, "completions/mean_length": 499.203125, "completions/mean_terminated_length": 434.75439453125, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 1.7351999999999999, "grad_norm": 0.038640473037958145, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57791727.0, "reward": 0.9362399578094482, "reward_std": 0.29871129989624023, "rewards/accuracy_reward": 0.484375, "rewards/brier_reward": 0.5287281274795532, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.859375, "rewards/mean_confidence_reward": 0.8492187261581421, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 939.0, "completions/mean_length": 490.9375, "completions/mean_terminated_length": 482.4762268066406, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 1.736, "grad_norm": 0.04978059604763985, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57838811.0, "reward": 0.9222360849380493, "reward_std": 0.3293876647949219, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.4694703221321106, "rewards/confidence_one_or_zero": 0.03125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.9167187213897705, "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": 0.078125, "completions/max_length": 1024.0, "completions/max_terminated_length": 963.0, "completions/mean_length": 652.59375, "completions/mean_terminated_length": 621.11865234375, "completions/min_length": 296.0, "completions/min_terminated_length": 296.0, "epoch": 1.7368000000000001, "grad_norm": 0.0478779636323452, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 57897777.0, "reward": 0.7348135709762573, "reward_std": 0.3478683531284332, "rewards/accuracy_reward": 0.203125, "rewards/brier_reward": 0.36024999618530273, "rewards/confidence_one_or_zero": 0.265625, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8134375214576721, "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": 997.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 489.34375, "completions/mean_terminated_length": 489.34375, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 1.7376, "grad_norm": 0.047614727169275284, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 57945399.0, "reward": 0.8752219676971436, "reward_std": 0.3764537274837494, "rewards/accuracy_reward": 0.3125, "rewards/brier_reward": 0.4379422068595886, "rewards/confidence_one_or_zero": 0.046875, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9001562595367432, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 1013.0, "completions/mean_length": 618.796875, "completions/mean_terminated_length": 598.8688354492188, "completions/min_length": 281.0, "completions/min_terminated_length": 281.0, "epoch": 1.7384, "grad_norm": 0.06373666226863861, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58000762.0, "reward": 0.9136977195739746, "reward_std": 0.4066515564918518, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.4992687404155731, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.7934374809265137, "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": 0.046875, "completions/max_length": 1024.0, "completions/max_terminated_length": 962.0, "completions/mean_length": 633.796875, "completions/mean_terminated_length": 614.6065063476562, "completions/min_length": 384.0, "completions/min_terminated_length": 384.0, "epoch": 1.7391999999999999, "grad_norm": 0.04816131666302681, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58058941.0, "reward": 1.053117275238037, "reward_std": 0.4204968810081482, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.637482762336731, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.844531238079071, "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": 873.0, "completions/max_terminated_length": 873.0, "completions/mean_length": 415.734375, "completions/mean_terminated_length": 415.734375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 1.74, "grad_norm": 0.0616181455552578, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58100828.0, "reward": 0.9278173446655273, "reward_std": 0.32828792929649353, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5275077819824219, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8317187428474426, "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": 0.09375, "completions/max_length": 1024.0, "completions/max_terminated_length": 959.0, "completions/mean_length": 545.578125, "completions/mean_terminated_length": 496.0862121582031, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 1.7408000000000001, "grad_norm": 0.06613653898239136, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58153601.0, "reward": 0.902114987373352, "reward_std": 0.3948025107383728, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.47610312700271606, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.8553124666213989, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 1011.0, "completions/mean_length": 497.328125, "completions/mean_terminated_length": 480.33868408203125, "completions/min_length": 162.0, "completions/min_terminated_length": 162.0, "epoch": 1.7416, "grad_norm": 0.05094151198863983, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58203174.0, "reward": 0.9353141188621521, "reward_std": 0.2723286747932434, "rewards/accuracy_reward": 0.34375, "rewards/brier_reward": 0.5581265687942505, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.788281261920929, "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": 0.109375, "completions/max_length": 1024.0, "completions/max_terminated_length": 968.0, "completions/mean_length": 514.375, "completions/mean_terminated_length": 451.78948974609375, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 1.7424, "grad_norm": 0.07396532595157623, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58252286.0, "reward": 0.9811336398124695, "reward_std": 0.37916305661201477, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.5560156106948853, "rewards/confidence_one_or_zero": 0.015625, "rewards/format_reward": 0.890625, "rewards/mean_confidence_reward": 0.8203125, "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": 953.0, "completions/max_terminated_length": 953.0, "completions/mean_length": 430.046875, "completions/mean_terminated_length": 430.046875, "completions/min_length": 174.0, "completions/min_terminated_length": 174.0, "epoch": 1.7431999999999999, "grad_norm": 0.06307395547628403, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58296305.0, "reward": 0.9414752125740051, "reward_std": 0.3878864049911499, "rewards/accuracy_reward": 0.375, "rewards/brier_reward": 0.5079483985900879, "rewards/confidence_one_or_zero": 0.09375, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.9001562595367432, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 932.0, "completions/mean_length": 470.453125, "completions/mean_terminated_length": 433.5500183105469, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 1.744, "grad_norm": 0.08836141973733902, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58343566.0, "reward": 1.0950236320495605, "reward_std": 0.37943288683891296, "rewards/accuracy_reward": 0.546875, "rewards/brier_reward": 0.7369202971458435, "rewards/confidence_one_or_zero": 0.1875, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7651562690734863, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 952.0, "completions/mean_length": 605.8125, "completions/mean_terminated_length": 592.3225708007812, "completions/min_length": 326.0, "completions/min_terminated_length": 326.0, "epoch": 1.7448000000000001, "grad_norm": 0.036798447370529175, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58400066.0, "reward": 0.8428078889846802, "reward_std": 0.2961253821849823, "rewards/accuracy_reward": 0.21875, "rewards/brier_reward": 0.49811404943466187, "rewards/confidence_one_or_zero": 0.140625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.750781238079071, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 916.0, "completions/mean_length": 449.0, "completions/mean_terminated_length": 430.45159912109375, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 1.7456, "grad_norm": 0.07004804909229279, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58444242.0, "reward": 1.234660267829895, "reward_std": 0.3581804037094116, "rewards/accuracy_reward": 0.703125, "rewards/brier_reward": 0.7974437475204468, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.8481249809265137, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 874.0, "completions/mean_length": 451.59375, "completions/mean_terminated_length": 433.1290283203125, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "epoch": 1.7464, "grad_norm": 0.08488323539495468, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58490168.0, "reward": 1.0476306676864624, "reward_std": 0.38288819789886475, "rewards/accuracy_reward": 0.515625, "rewards/brier_reward": 0.6265093088150024, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.953125, "rewards/mean_confidence_reward": 0.81640625, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 888.0, "completions/mean_length": 595.078125, "completions/mean_terminated_length": 566.4833374023438, "completions/min_length": 342.0, "completions/min_terminated_length": 342.0, "epoch": 1.7471999999999999, "grad_norm": 0.04653465747833252, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58546317.0, "reward": 0.8278430700302124, "reward_std": 0.3453921377658844, "rewards/accuracy_reward": 0.296875, "rewards/brier_reward": 0.45255935192108154, "rewards/confidence_one_or_zero": 0.109375, "rewards/format_reward": 0.90625, "rewards/mean_confidence_reward": 0.7768750190734863, "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": 0.0625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1022.0, "completions/mean_length": 575.328125, "completions/mean_terminated_length": 545.4166870117188, "completions/min_length": 238.0, "completions/min_terminated_length": 238.0, "epoch": 1.748, "grad_norm": 0.04307619854807854, "learning_rate": 1e-06, "loss": 0.0, "num_tokens": 58599682.0, "reward": 0.9446532726287842, "reward_std": 0.35438278317451477, "rewards/accuracy_reward": 0.421875, "rewards/brier_reward": 0.5299296975135803, "rewards/confidence_one_or_zero": 0.0625, "rewards/format_reward": 0.9375, "rewards/mean_confidence_reward": 0.8517187237739563, "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": 0.0, "completions/max_length": 703.0, "completions/max_terminated_length": 703.0, "completions/mean_length": 331.015625, "completions/mean_terminated_length": 331.015625, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "epoch": 1.7488000000000001, "grad_norm": 0.11811847239732742, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58637107.0, "reward": 1.2873940467834473, "reward_std": 0.3478180170059204, "rewards/accuracy_reward": 0.78125, "rewards/brier_reward": 0.7935359477996826, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 1.0, "rewards/mean_confidence_reward": 0.95640629529953, "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": 0.015625, "completions/max_length": 1024.0, "completions/max_terminated_length": 1007.0, "completions/mean_length": 548.3125, "completions/mean_terminated_length": 540.761962890625, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 1.7496, "grad_norm": 0.043763358145952225, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58688871.0, "reward": 1.1963634490966797, "reward_std": 0.27715304493904114, "rewards/accuracy_reward": 0.65625, "rewards/brier_reward": 0.7520999908447266, "rewards/confidence_one_or_zero": 0.125, "rewards/format_reward": 0.984375, "rewards/mean_confidence_reward": 0.8778125047683716, "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": 0.03125, "completions/max_length": 1024.0, "completions/max_terminated_length": 907.0, "completions/mean_length": 503.953125, "completions/mean_terminated_length": 487.1773986816406, "completions/min_length": 265.0, "completions/min_terminated_length": 265.0, "epoch": 1.7504, "grad_norm": 0.06587336212396622, "learning_rate": 1e-06, "loss": -0.0, "num_tokens": 58737780.0, "reward": 0.9937118291854858, "reward_std": 0.3911515176296234, "rewards/accuracy_reward": 0.390625, "rewards/brier_reward": 0.6280468702316284, "rewards/confidence_one_or_zero": 0.15625, "rewards/format_reward": 0.96875, "rewards/mean_confidence_reward": 0.7640625238418579, "step": 2188 }, { "epoch": 1.7504, "step": 2188, "total_flos": 0.0, "train_loss": 4.737286882090231e-10, "train_runtime": 30658.6653, "train_samples_per_second": 4.567, "train_steps_per_second": 0.071 } ], "logging_steps": 1, "max_steps": 2188, "num_input_tokens_seen": 58737780, "num_train_epochs": 2, "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 }