{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 50,
  "global_step": 63,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "entropy": 2.5553497850894926,
      "epoch": 0.08,
      "grad_norm": 357.8926696777344,
      "learning_rate": 2.8571428571428573e-06,
      "logits/chosen": 0.41187580737484436,
      "logits/rejected": 0.43087167594886183,
      "logps/chosen": -157.8165241241455,
      "logps/rejected": -150.432616519928,
      "loss": 0.6890408515930175,
      "mean_token_accuracy": 0.45499045923352244,
      "num_tokens": 63280.0,
      "rewards/accuracies": 0.55625,
      "rewards/chosen": 0.18442693948745728,
      "rewards/margins": 0.05980904102325439,
      "rewards/rejected": 0.12461789846420288,
      "step": 5
    },
    {
      "entropy": 2.585687577724457,
      "epoch": 0.16,
      "grad_norm": 550.5374145507812,
      "learning_rate": 4.984280524733107e-06,
      "logits/chosen": 0.23282768258015127,
      "logits/rejected": 0.41657135452577226,
      "logps/chosen": -160.46632223129274,
      "logps/rejected": -155.05524663925172,
      "loss": 1.0951576232910156,
      "mean_token_accuracy": 0.4666588671505451,
      "num_tokens": 128751.0,
      "rewards/accuracies": 0.46875,
      "rewards/chosen": 0.04196357727050781,
      "rewards/margins": -0.13510316610336304,
      "rewards/rejected": 0.17706674337387085,
      "step": 10
    },
    {
      "entropy": 2.4887936890125273,
      "epoch": 0.24,
      "grad_norm": 544.36474609375,
      "learning_rate": 4.809698831278217e-06,
      "logits/chosen": 0.19960332682723014,
      "logits/rejected": 0.25714549020005323,
      "logps/chosen": -151.403657245636,
      "logps/rejected": -163.51016159057616,
      "loss": 1.1309956550598144,
      "mean_token_accuracy": 0.4729401096701622,
      "num_tokens": 194142.0,
      "rewards/accuracies": 0.45625,
      "rewards/chosen": -0.6780350089073182,
      "rewards/margins": 0.11501268744468689,
      "rewards/rejected": -0.793047696352005,
      "step": 15
    },
    {
      "entropy": 2.543038326501846,
      "epoch": 0.32,
      "grad_norm": 545.773193359375,
      "learning_rate": 4.454578706170075e-06,
      "logits/chosen": 0.10757563983449289,
      "logits/rejected": 0.2191312620805499,
      "logps/chosen": -134.6160714149475,
      "logps/rejected": -134.7640838623047,
      "loss": 1.2686697959899902,
      "mean_token_accuracy": 0.46390472874045374,
      "num_tokens": 254871.0,
      "rewards/accuracies": 0.54375,
      "rewards/chosen": -1.045345139503479,
      "rewards/margins": 0.4233867168426514,
      "rewards/rejected": -1.4687318563461305,
      "step": 20
    },
    {
      "entropy": 2.4839668542146685,
      "epoch": 0.4,
      "grad_norm": 609.0315551757812,
      "learning_rate": 3.946678240449515e-06,
      "logits/chosen": 0.23401674829751978,
      "logits/rejected": 0.22746947777638224,
      "logps/chosen": -144.78962602615357,
      "logps/rejected": -136.76443700790406,
      "loss": 1.4144909858703614,
      "mean_token_accuracy": 0.45985892340540885,
      "num_tokens": 317752.0,
      "rewards/accuracies": 0.5625,
      "rewards/chosen": -0.8229869723320007,
      "rewards/margins": 0.27719623446464536,
      "rewards/rejected": -1.1001832067966462,
      "step": 25
    },
    {
      "entropy": 2.4549559235572813,
      "epoch": 0.48,
      "grad_norm": 685.6503295898438,
      "learning_rate": 3.3256976548879183e-06,
      "logits/chosen": 0.21252182170691483,
      "logits/rejected": 0.27063995156819315,
      "logps/chosen": -148.57344274520875,
      "logps/rejected": -146.81006059646606,
      "loss": 1.6434017181396485,
      "mean_token_accuracy": 0.454254624247551,
      "num_tokens": 387452.0,
      "rewards/accuracies": 0.4625,
      "rewards/chosen": -0.23523921966552735,
      "rewards/margins": 0.0028411030769348146,
      "rewards/rejected": -0.23808032274246216,
      "step": 30
    },
    {
      "entropy": 2.4848766714334487,
      "epoch": 0.56,
      "grad_norm": 572.0255737304688,
      "learning_rate": 2.6401761180929798e-06,
      "logits/chosen": 0.1604402064572174,
      "logits/rejected": 0.23192453424270135,
      "logps/chosen": -178.91495656967163,
      "logps/rejected": -171.79455051422119,
      "loss": 1.2930749893188476,
      "mean_token_accuracy": 0.4774269789457321,
      "num_tokens": 454873.0,
      "rewards/accuracies": 0.54375,
      "rewards/chosen": 0.2875680088996887,
      "rewards/margins": 0.7235915064811707,
      "rewards/rejected": -0.43602349758148196,
      "step": 35
    },
    {
      "entropy": 2.4956657856702806,
      "epoch": 0.64,
      "grad_norm": 581.42724609375,
      "learning_rate": 1.9436976651092143e-06,
      "logits/chosen": 0.26337065405985355,
      "logits/rejected": 0.36300255439029444,
      "logps/chosen": -157.99055194854736,
      "logps/rejected": -159.0982545375824,
      "loss": 1.5171399116516113,
      "mean_token_accuracy": 0.46507507339119913,
      "num_tokens": 518527.0,
      "rewards/accuracies": 0.5,
      "rewards/chosen": -0.061191481351852414,
      "rewards/margins": 0.1269888162612915,
      "rewards/rejected": -0.18818029761314392,
      "step": 40
    },
    {
      "entropy": 2.434966567158699,
      "epoch": 0.72,
      "grad_norm": 432.56536865234375,
      "learning_rate": 1.2907027822369006e-06,
      "logits/chosen": 0.268132887364752,
      "logits/rejected": 0.30894473259910554,
      "logps/chosen": -154.1576229095459,
      "logps/rejected": -159.85538330078126,
      "loss": 1.3509400367736817,
      "mean_token_accuracy": 0.4740308240056038,
      "num_tokens": 586496.0,
      "rewards/accuracies": 0.5375,
      "rewards/chosen": 0.02395761013031006,
      "rewards/margins": 0.8281095027923584,
      "rewards/rejected": -0.8041518926620483,
      "step": 45
    },
    {
      "entropy": 2.4690693795681,
      "epoch": 0.8,
      "grad_norm": 504.9461364746094,
      "learning_rate": 7.322330470336314e-07,
      "logits/chosen": 0.27772643533489566,
      "logits/rejected": 0.36052606081661537,
      "logps/chosen": -164.17142915725708,
      "logps/rejected": -163.73542156219483,
      "loss": 1.4861845970153809,
      "mean_token_accuracy": 0.4653610520064831,
      "num_tokens": 659345.0,
      "rewards/accuracies": 0.55625,
      "rewards/chosen": -0.0009298324584960938,
      "rewards/margins": 0.31063854694366455,
      "rewards/rejected": -0.31156837940216064,
      "step": 50
    },
    {
      "epoch": 0.8,
      "eval_entropy": 2.536505832195282,
      "eval_logits/chosen": 0.2893901140152919,
      "eval_logits/rejected": 0.34482142702810936,
      "eval_logps/chosen": -161.61729994201661,
      "eval_logps/rejected": -153.7386490020752,
      "eval_loss": 1.3951599597930908,
      "eval_mean_token_accuracy": 0.46117462730407716,
      "eval_num_tokens": 659345.0,
      "eval_rewards/accuracies": 0.539,
      "eval_rewards/chosen": -0.16949110031127929,
      "eval_rewards/margins": 0.413674391746521,
      "eval_rewards/rejected": -0.5831654920578003,
      "eval_runtime": 204.2784,
      "eval_samples_per_second": 4.895,
      "eval_steps_per_second": 1.224,
      "step": 50
    },
    {
      "entropy": 2.5048249810934067,
      "epoch": 0.88,
      "grad_norm": 538.51953125,
      "learning_rate": 3.119414452281158e-07,
      "logits/chosen": 0.31241473497619054,
      "logits/rejected": 0.4336116565754852,
      "logps/chosen": -148.8257155418396,
      "logps/rejected": -143.5335810661316,
      "loss": 1.5545385360717774,
      "mean_token_accuracy": 0.47073024958372117,
      "num_tokens": 731475.0,
      "rewards/accuracies": 0.53125,
      "rewards/chosen": -0.08134242296218872,
      "rewards/margins": 0.31841338872909547,
      "rewards/rejected": -0.39975581169128416,
      "step": 55
    },
    {
      "entropy": 2.4619424283504485,
      "epoch": 0.96,
      "grad_norm": 439.4546813964844,
      "learning_rate": 6.268021954544095e-08,
      "logits/chosen": 0.26771219837052634,
      "logits/rejected": 0.3957730213546686,
      "logps/chosen": -157.09008331298827,
      "logps/rejected": -164.03927259445192,
      "loss": 1.2014739990234375,
      "mean_token_accuracy": 0.47436425387859343,
      "num_tokens": 797757.0,
      "rewards/accuracies": 0.525,
      "rewards/chosen": 0.21749011874198915,
      "rewards/margins": 0.5742806613445282,
      "rewards/rejected": -0.35679054260253906,
      "step": 60
    },
    {
      "epoch": 1.0,
      "eval_entropy": 2.5392660574913024,
      "eval_logits/chosen": 0.2875409764336183,
      "eval_logits/rejected": 0.34302489192485985,
      "eval_logps/chosen": -161.68841091918947,
      "eval_logps/rejected": -153.80732011413573,
      "eval_loss": 1.3833612203598022,
      "eval_mean_token_accuracy": 0.46109312522411344,
      "eval_num_tokens": 829784.0,
      "eval_rewards/accuracies": 0.543,
      "eval_rewards/chosen": -0.24060248374938964,
      "eval_rewards/margins": 0.4112340784072876,
      "eval_rewards/rejected": -0.6518365621566773,
      "eval_runtime": 204.0071,
      "eval_samples_per_second": 4.902,
      "eval_steps_per_second": 1.225,
      "step": 63
    }
  ],
  "logging_steps": 5,
  "max_steps": 63,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3267470732328960.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}