{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.9790660225442833,
  "eval_steps": 50,
  "global_step": 57,
  "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": 0.045703125,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1497.8,
      "completions/mean_length": 521.414453125,
      "completions/mean_terminated_length": 472.765625,
      "completions/min_length": 74.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.2576489533011272,
      "grad_norm": 0.1361742913722992,
      "learning_rate": 1e-06,
      "loss": 0.0562,
      "num_tokens": 4038541.0,
      "reward": 0.4758795499801636,
      "reward_std": 0.4125692009925842,
      "rewards/accuracy_reward": 0.064453125,
      "rewards/brier_reward": 0.2697267413139343,
      "rewards/confidence_one_or_zero": 0.06171875,
      "rewards/format_reward": 0.617578125,
      "rewards/mean_confidence_reward": 0.500839501619339,
      "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": 0.04609375,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1505.2,
      "completions/mean_length": 523.770703125,
      "completions/mean_terminated_length": 474.846337890625,
      "completions/min_length": 71.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.5152979066022544,
      "grad_norm": 0.0030647902749478817,
      "learning_rate": 1e-06,
      "loss": 0.0583,
      "num_tokens": 8106450.0,
      "reward": 0.5272839426994324,
      "reward_std": 0.40863736867904665,
      "rewards/accuracy_reward": 0.07109375,
      "rewards/brier_reward": 0.3018323004245758,
      "rewards/confidence_one_or_zero": 0.078125,
      "rewards/format_reward": 0.681640625,
      "rewards/mean_confidence_reward": 0.5452994704246521,
      "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": 0.039453125,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1519.0,
      "completions/mean_length": 499.593359375,
      "completions/mean_terminated_length": 457.06886596679686,
      "completions/min_length": 67.2,
      "completions/min_terminated_length": 67.2,
      "epoch": 0.7729468599033816,
      "grad_norm": 0.002637381898239255,
      "learning_rate": 1e-06,
      "loss": 0.047,
      "num_tokens": 12081497.0,
      "reward": 0.6300971508026123,
      "reward_std": 0.3623956501483917,
      "rewards/accuracy_reward": 0.086328125,
      "rewards/brier_reward": 0.35863040685653685,
      "rewards/confidence_one_or_zero": 0.071484375,
      "rewards/format_reward": 0.815234375,
      "rewards/mean_confidence_reward": 0.6512776494026185,
      "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": 0.0331206995412844,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1507.6,
      "completions/mean_length": 500.6675964355469,
      "completions/mean_terminated_length": 465.1861938476562,
      "completions/min_length": 76.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 1.0515297906602254,
      "grad_norm": 0.001991912489756942,
      "learning_rate": 1e-06,
      "loss": 0.0578,
      "num_tokens": 16048198.0,
      "reward": 0.7004695773124695,
      "reward_std": 0.32459375262260437,
      "rewards/accuracy_reward": 0.096875,
      "rewards/brier_reward": 0.40406264662742614,
      "rewards/confidence_one_or_zero": 0.078125,
      "rewards/format_reward": 0.9,
      "rewards/mean_confidence_reward": 0.6946446537971497,
      "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": 0.02890625,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1522.6,
      "completions/mean_length": 469.98359375,
      "completions/mean_terminated_length": 438.2403137207031,
      "completions/min_length": 66.6,
      "completions/min_terminated_length": 66.6,
      "epoch": 1.3091787439613527,
      "grad_norm": 0.32453644275665283,
      "learning_rate": 1e-06,
      "loss": 0.0493,
      "num_tokens": 19948636.0,
      "reward": 0.7361393928527832,
      "reward_std": 0.315061616897583,
      "rewards/accuracy_reward": 0.107421875,
      "rewards/brier_reward": 0.4398553490638733,
      "rewards/confidence_one_or_zero": 0.07734375,
      "rewards/format_reward": 0.925,
      "rewards/mean_confidence_reward": 0.6970306396484375,
      "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": 0.025,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1524.8,
      "completions/mean_length": 444.559765625,
      "completions/mean_terminated_length": 416.5305541992187,
      "completions/min_length": 77.6,
      "completions/min_terminated_length": 77.6,
      "epoch": 1.5668276972624797,
      "grad_norm": 0.00270429952070117,
      "learning_rate": 1e-06,
      "loss": 0.0415,
      "num_tokens": 23791453.0,
      "reward": 0.7462241291999817,
      "reward_std": 0.30905682444572447,
      "rewards/accuracy_reward": 0.110546875,
      "rewards/brier_reward": 0.4447904586791992,
      "rewards/confidence_one_or_zero": 0.06484375,
      "rewards/format_reward": 0.937109375,
      "rewards/mean_confidence_reward": 0.7072680354118347,
      "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": 0.025390625,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1511.2,
      "completions/mean_length": 425.152734375,
      "completions/mean_terminated_length": 396.2151611328125,
      "completions/min_length": 50.8,
      "completions/min_terminated_length": 50.8,
      "epoch": 1.8244766505636072,
      "grad_norm": 0.002345304237678647,
      "learning_rate": 1e-06,
      "loss": 0.0458,
      "num_tokens": 27576596.0,
      "reward": 0.7759710311889648,
      "reward_std": 0.30494046211242676,
      "rewards/accuracy_reward": 0.1234375,
      "rewards/brier_reward": 0.4824093222618103,
      "rewards/confidence_one_or_zero": 0.06796875,
      "rewards/format_reward": 0.94609375,
      "rewards/mean_confidence_reward": 0.681143045425415,
      "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": 0.01511611238532109,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1485.4,
      "completions/mean_length": 392.34654541015624,
      "completions/mean_terminated_length": 374.88779907226564,
      "completions/min_length": 82.6,
      "completions/min_terminated_length": 82.6,
      "epoch": 2.103059581320451,
      "grad_norm": 0.33194369077682495,
      "learning_rate": 1e-06,
      "loss": 0.0346,
      "num_tokens": 31277252.0,
      "reward": 0.8171104550361633,
      "reward_std": 0.3093654751777649,
      "rewards/accuracy_reward": 0.157421875,
      "rewards/brier_reward": 0.5166412889957428,
      "rewards/confidence_one_or_zero": 0.06953125,
      "rewards/format_reward": 0.96015625,
      "rewards/mean_confidence_reward": 0.6755032181739807,
      "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": 0.013671875,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1418.8,
      "completions/mean_length": 357.22265625,
      "completions/mean_terminated_length": 340.8806396484375,
      "completions/min_length": 74.4,
      "completions/min_terminated_length": 74.4,
      "epoch": 2.360708534621578,
      "grad_norm": 0.00267302175052464,
      "learning_rate": 1e-06,
      "loss": 0.035,
      "num_tokens": 34897222.0,
      "reward": 0.8488460063934327,
      "reward_std": 0.31457024812698364,
      "rewards/accuracy_reward": 0.181640625,
      "rewards/brier_reward": 0.5469092965126038,
      "rewards/confidence_one_or_zero": 0.066015625,
      "rewards/format_reward": 0.969140625,
      "rewards/mean_confidence_reward": 0.6639328360557556,
      "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": 0.013671875,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1464.4,
      "completions/mean_length": 345.26796875,
      "completions/mean_terminated_length": 328.7436828613281,
      "completions/min_length": 70.8,
      "completions/min_terminated_length": 70.8,
      "epoch": 2.6183574879227054,
      "grad_norm": 0.003008283441886306,
      "learning_rate": 1e-06,
      "loss": 0.0415,
      "num_tokens": 38456532.0,
      "reward": 0.8908233880996704,
      "reward_std": 0.3146331191062927,
      "rewards/accuracy_reward": 0.21796875,
      "rewards/brier_reward": 0.589457881450653,
      "rewards/confidence_one_or_zero": 0.055078125,
      "rewards/format_reward": 0.97421875,
      "rewards/mean_confidence_reward": 0.6358984112739563,
      "step": 50
    },
    {
      "epoch": 2.6183574879227054,
      "eval_completions/clipped_ratio": 0.009057971014492754,
      "eval_completions/max_length": 788.3478260869565,
      "eval_completions/max_terminated_length": 733.0144927536232,
      "eval_completions/mean_length": 323.0208333333333,
      "eval_completions/mean_terminated_length": 312.0772527058919,
      "eval_completions/min_length": 131.19565217391303,
      "eval_completions/min_terminated_length": 131.19565217391303,
      "eval_loss": 0.0,
      "eval_num_tokens": 38456532.0,
      "eval_reward": 0.9158611794312795,
      "eval_reward_std": 0.3136252526571785,
      "eval_rewards/accuracy_reward": 0.24094202898550723,
      "eval_rewards/brier_reward": 0.6107065263001815,
      "eval_rewards/confidence_one_or_zero": 0.050724637681159424,
      "eval_rewards/format_reward": 0.980072463768116,
      "eval_rewards/mean_confidence_reward": 0.6292391296314157,
      "eval_runtime": 570.9473,
      "eval_samples_per_second": 0.242,
      "eval_steps_per_second": 0.032,
      "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": 0.01328125,
      "completions/max_length": 1536.0,
      "completions/max_terminated_length": 1483.8,
      "completions/mean_length": 309.042578125,
      "completions/mean_terminated_length": 292.53196411132814,
      "completions/min_length": 71.8,
      "completions/min_terminated_length": 71.8,
      "epoch": 2.8760064412238324,
      "grad_norm": 0.0024981796741485596,
      "learning_rate": 1e-06,
      "loss": 0.0362,
      "num_tokens": 41974297.0,
      "reward": 0.936638355255127,
      "reward_std": 0.3146317660808563,
      "rewards/accuracy_reward": 0.264453125,
      "rewards/brier_reward": 0.6279629111289978,
      "rewards/confidence_one_or_zero": 0.05390625,
      "rewards/format_reward": 0.980859375,
      "rewards/mean_confidence_reward": 0.6118968725204468,
      "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": 0.0022935779816513624,
      "completions/max_length": 1418.0,
      "completions/max_terminated_length": 1323.5,
      "completions/mean_length": 279.01670837402344,
      "completions/mean_terminated_length": 276.12876892089844,
      "completions/min_length": 71.5,
      "completions/min_terminated_length": 71.5,
      "epoch": 2.9790660225442833,
      "num_tokens": 43354056.0,
      "reward": 0.970618486404419,
      "reward_std": 0.30738210678100586,
      "rewards/accuracy_reward": 0.3017578125,
      "rewards/brier_reward": 0.6502200961112976,
      "rewards/confidence_one_or_zero": 0.041015625,
      "rewards/format_reward": 0.9892578125,
      "rewards/mean_confidence_reward": 0.6022265553474426,
      "step": 57,
      "total_flos": 0.0,
      "train_loss": 0.04462047953877533,
      "train_runtime": 11303.71,
      "train_samples_per_second": 0.33,
      "train_steps_per_second": 0.005
    }
  ],
  "logging_steps": 5,
  "max_steps": 57,
  "num_input_tokens_seen": 43354056,
  "num_train_epochs": 3,
  "save_steps": 60,
  "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": 8,
  "trial_name": null,
  "trial_params": null
}