{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9757287474082205,
  "eval_steps": 1000,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.24,
      "learning_rate": 4.30097988740375e-05,
      "logits/chosen": 0.4760148227214813,
      "logits/rejected": 0.18961936235427856,
      "logps/chosen": -138.3477783203125,
      "logps/rejected": -107.43299865722656,
      "loss": 0.0534,
      "rewards/accuracies": 0.9904999732971191,
      "rewards/chosen": -1.208303689956665,
      "rewards/margins": 6.7370734214782715,
      "rewards/rejected": -7.945377349853516,
      "step": 1000
    },
    {
      "epoch": 0.24,
      "eval_logits/chosen": 0.5734850168228149,
      "eval_logits/rejected": 0.30642059445381165,
      "eval_logps/chosen": -140.49411010742188,
      "eval_logps/rejected": -130.59207153320312,
      "eval_loss": 0.021729281172156334,
      "eval_rewards/accuracies": 0.994514524936676,
      "eval_rewards/chosen": -1.671390414237976,
      "eval_rewards/margins": 8.56452751159668,
      "eval_rewards/rejected": -10.235918045043945,
      "eval_runtime": 379.8081,
      "eval_samples_per_second": 9.597,
      "eval_steps_per_second": 4.8,
      "step": 1000
    },
    {
      "epoch": 0.49,
      "learning_rate": 2.594822843866259e-05,
      "logits/chosen": 0.9377780556678772,
      "logits/rejected": 0.5601473450660706,
      "logps/chosen": -137.52035522460938,
      "logps/rejected": -132.42645263671875,
      "loss": 0.0182,
      "rewards/accuracies": 0.9955000281333923,
      "rewards/chosen": -1.4818373918533325,
      "rewards/margins": 8.971912384033203,
      "rewards/rejected": -10.453749656677246,
      "step": 2000
    },
    {
      "epoch": 0.49,
      "eval_logits/chosen": 1.070937156677246,
      "eval_logits/rejected": 0.6229708194732666,
      "eval_logps/chosen": -139.24871826171875,
      "eval_logps/rejected": -137.83485412597656,
      "eval_loss": 0.01748650148510933,
      "eval_rewards/accuracies": 0.995063066482544,
      "eval_rewards/chosen": -1.5468522310256958,
      "eval_rewards/margins": 9.413342475891113,
      "eval_rewards/rejected": -10.96019458770752,
      "eval_runtime": 382.2871,
      "eval_samples_per_second": 9.535,
      "eval_steps_per_second": 4.769,
      "step": 2000
    },
    {
      "epoch": 0.73,
      "learning_rate": 8.35639340331897e-06,
      "logits/chosen": 1.0557544231414795,
      "logits/rejected": 0.6803003549575806,
      "logps/chosen": -137.27392578125,
      "logps/rejected": -142.61241149902344,
      "loss": 0.0162,
      "rewards/accuracies": 0.9959999918937683,
      "rewards/chosen": -1.7070084810256958,
      "rewards/margins": 9.696061134338379,
      "rewards/rejected": -11.403070449829102,
      "step": 3000
    },
    {
      "epoch": 0.73,
      "eval_logits/chosen": 1.0047662258148193,
      "eval_logits/rejected": 0.6324675679206848,
      "eval_logps/chosen": -139.297607421875,
      "eval_logps/rejected": -142.67721557617188,
      "eval_loss": 0.017071055248379707,
      "eval_rewards/accuracies": 0.9961601495742798,
      "eval_rewards/chosen": -1.5517414808273315,
      "eval_rewards/margins": 9.89268970489502,
      "eval_rewards/rejected": -11.444432258605957,
      "eval_runtime": 382.4643,
      "eval_samples_per_second": 9.53,
      "eval_steps_per_second": 4.766,
      "step": 3000
    },
    {
      "epoch": 0.98,
      "learning_rate": 7.193097375108037e-08,
      "logits/chosen": 0.9620022773742676,
      "logits/rejected": 0.5997971892356873,
      "logps/chosen": -140.70065307617188,
      "logps/rejected": -143.5229949951172,
      "loss": 0.0154,
      "rewards/accuracies": 0.9947500228881836,
      "rewards/chosen": -1.644036054611206,
      "rewards/margins": 9.865877151489258,
      "rewards/rejected": -11.509913444519043,
      "step": 4000
    },
    {
      "epoch": 0.98,
      "eval_logits/chosen": 0.9728695154190063,
      "eval_logits/rejected": 0.6051246523857117,
      "eval_logps/chosen": -139.5214080810547,
      "eval_logps/rejected": -142.23641967773438,
      "eval_loss": 0.016766143962740898,
      "eval_rewards/accuracies": 0.9956116080284119,
      "eval_rewards/chosen": -1.5741225481033325,
      "eval_rewards/margins": 9.8262300491333,
      "eval_rewards/rejected": -11.40035343170166,
      "eval_runtime": 383.0087,
      "eval_samples_per_second": 9.517,
      "eval_steps_per_second": 4.76,
      "step": 4000
    }
  ],
  "logging_steps": 1000,
  "max_steps": 4099,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "total_flos": 1.7962050589084877e+17,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}