{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.96,
  "eval_steps": 500,
  "global_step": 24,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "clip_ratio": 0.0,
      "completion_length": 532.6341304779053,
      "epoch": 0.32,
      "grad_norm": 0.9071540047756161,
      "kl": 0.0,
      "learning_rate": 3.3333333333333333e-06,
      "loss": -0.006,
      "num_tokens": 1244103.0,
      "reward": 0.581380233168602,
      "reward_std": 0.17345089931041002,
      "rewards/compiled_reward": 0.5813802070915699,
      "step": 2
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 592.2552280426025,
      "epoch": 0.64,
      "grad_norm": 1.1007144051464353,
      "kl": 0.009278297424316406,
      "learning_rate": 1e-05,
      "loss": 0.0884,
      "num_tokens": 2533075.0,
      "reward": 1.2021484710276127,
      "reward_std": 0.7038572877645493,
      "rewards/compiled_reward": 1.2021484412252903,
      "step": 4
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 489.96876335144043,
      "epoch": 0.96,
      "grad_norm": 0.9559919963618165,
      "kl": 0.06231689453125,
      "learning_rate": 9.777864028930705e-06,
      "loss": 0.0432,
      "num_tokens": 3786199.0,
      "reward": 1.9729818180203438,
      "reward_std": 0.8437984921038151,
      "rewards/compiled_reward": 1.9729817658662796,
      "step": 6
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 514.4049606323242,
      "epoch": 1.32,
      "grad_norm": 0.8994591128856398,
      "kl": 0.0983123779296875,
      "learning_rate": 9.131193871579975e-06,
      "loss": 0.062,
      "num_tokens": 5038070.0,
      "reward": 2.1230469569563866,
      "reward_std": 0.8294221684336662,
      "rewards/compiled_reward": 2.123046852648258,
      "step": 8
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 566.2734565734863,
      "epoch": 1.6400000000000001,
      "grad_norm": 0.7238876816698709,
      "kl": 0.1275634765625,
      "learning_rate": 8.117449009293668e-06,
      "loss": 0.0575,
      "num_tokens": 6317864.0,
      "reward": 2.7200521528720856,
      "reward_std": 0.5833112373948097,
      "rewards/compiled_reward": 2.7200520783662796,
      "step": 10
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 539.0247611999512,
      "epoch": 1.96,
      "grad_norm": 0.6748076119984399,
      "kl": 0.145477294921875,
      "learning_rate": 6.8267051218319766e-06,
      "loss": 0.0371,
      "num_tokens": 7565020.0,
      "reward": 2.750000074505806,
      "reward_std": 0.40718091931194067,
      "rewards/compiled_reward": 2.750000014901161,
      "step": 12
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 525.386739730835,
      "epoch": 2.32,
      "grad_norm": 12.587918069852307,
      "kl": 0.261199951171875,
      "learning_rate": 5.373650467932122e-06,
      "loss": 0.0467,
      "num_tokens": 8817637.0,
      "reward": 2.9186198711395264,
      "reward_std": 0.32051649037748575,
      "rewards/compiled_reward": 2.9186198115348816,
      "step": 14
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 517.9622592926025,
      "epoch": 2.64,
      "grad_norm": 0.7052158461678741,
      "kl": 0.16070556640625,
      "learning_rate": 3.887395330218429e-06,
      "loss": 0.0419,
      "num_tokens": 10051840.0,
      "reward": 2.9300131052732468,
      "reward_std": 0.4993608547374606,
      "rewards/compiled_reward": 2.9300130009651184,
      "step": 16
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 467.34377098083496,
      "epoch": 2.96,
      "grad_norm": 0.8429441508712391,
      "kl": 0.166748046875,
      "learning_rate": 2.5000000000000015e-06,
      "loss": 0.0523,
      "num_tokens": 11267024.0,
      "reward": 2.7708334028720856,
      "reward_std": 0.7680677343159914,
      "rewards/compiled_reward": 2.7708333283662796,
      "step": 18
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 504.5286636352539,
      "epoch": 3.32,
      "grad_norm": 0.7525132694002632,
      "kl": 0.156494140625,
      "learning_rate": 1.3347406408508695e-06,
      "loss": 0.0645,
      "num_tokens": 12512150.0,
      "reward": 2.5833334028720856,
      "reward_std": 0.852101992815733,
      "rewards/compiled_reward": 2.583333343267441,
      "step": 20
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 489.8724060058594,
      "epoch": 3.64,
      "grad_norm": 0.7232325174102834,
      "kl": 0.17425537109375,
      "learning_rate": 4.951556604879049e-07,
      "loss": 0.058,
      "num_tokens": 13725860.0,
      "reward": 2.7301433235406876,
      "reward_std": 0.6854851758107543,
      "rewards/compiled_reward": 2.730143219232559,
      "step": 22
    },
    {
      "clip_ratio": 0.0,
      "completion_length": 480.85418128967285,
      "epoch": 3.96,
      "grad_norm": 0.8329086910965117,
      "kl": 0.18341064453125,
      "learning_rate": 5.584586887435739e-08,
      "loss": 0.0553,
      "num_tokens": 14944987.0,
      "reward": 2.772135466337204,
      "reward_std": 0.7318777162581682,
      "rewards/compiled_reward": 2.7721354216337204,
      "step": 24
    }
  ],
  "logging_steps": 2,
  "max_steps": 24,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 30,
  "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": 16,
  "trial_name": null,
  "trial_params": null
}