{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9874476987447699,
  "eval_steps": 500,
  "global_step": 59,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.016736401673640166,
      "grad_norm": 73.60199226633047,
      "learning_rate": 1.6666666666666664e-08,
      "logits/chosen": -2.3752830028533936,
      "logits/rejected": -2.3043034076690674,
      "logps/chosen": -249.16665649414062,
      "logps/pi_response": -145.31784057617188,
      "logps/ref_response": -145.31784057617188,
      "logps/rejected": -564.48095703125,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.16736401673640167,
      "grad_norm": 83.89151656409592,
      "learning_rate": 9.860114570402053e-08,
      "logits/chosen": -2.3579108715057373,
      "logits/rejected": -2.27597713470459,
      "logps/chosen": -267.1168518066406,
      "logps/pi_response": -122.99234008789062,
      "logps/ref_response": -122.98355102539062,
      "logps/rejected": -573.5260009765625,
      "loss": 0.6848,
      "rewards/accuracies": 0.5555555820465088,
      "rewards/chosen": -0.009232524782419205,
      "rewards/margins": 0.018979281187057495,
      "rewards/rejected": -0.0282118059694767,
      "step": 10
    },
    {
      "epoch": 0.33472803347280333,
      "grad_norm": 89.4856281132791,
      "learning_rate": 8.374915007591053e-08,
      "logits/chosen": -2.3042211532592773,
      "logits/rejected": -2.238210439682007,
      "logps/chosen": -305.8570251464844,
      "logps/pi_response": -132.1517333984375,
      "logps/ref_response": -130.00250244140625,
      "logps/rejected": -596.7067260742188,
      "loss": 0.6159,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.20397000014781952,
      "rewards/margins": 0.2074459046125412,
      "rewards/rejected": -0.4114159643650055,
      "step": 20
    },
    {
      "epoch": 0.502092050209205,
      "grad_norm": 38.21416099960407,
      "learning_rate": 5.738232820012406e-08,
      "logits/chosen": -2.2530503273010254,
      "logits/rejected": -2.1885294914245605,
      "logps/chosen": -311.74725341796875,
      "logps/pi_response": -127.53953552246094,
      "logps/ref_response": -124.12480163574219,
      "logps/rejected": -635.48291015625,
      "loss": 0.5539,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.32179349660873413,
      "rewards/margins": 0.5445557236671448,
      "rewards/rejected": -0.8663492202758789,
      "step": 30
    },
    {
      "epoch": 0.6694560669456067,
      "grad_norm": 34.96128217449131,
      "learning_rate": 2.8496739886173992e-08,
      "logits/chosen": -2.2476882934570312,
      "logits/rejected": -2.15933895111084,
      "logps/chosen": -313.03948974609375,
      "logps/pi_response": -147.4397430419922,
      "logps/ref_response": -138.77117919921875,
      "logps/rejected": -692.9878540039062,
      "loss": 0.5112,
      "rewards/accuracies": 0.768750011920929,
      "rewards/chosen": -0.43414002656936646,
      "rewards/margins": 0.7952297329902649,
      "rewards/rejected": -1.2293697595596313,
      "step": 40
    },
    {
      "epoch": 0.8368200836820083,
      "grad_norm": 33.66203802745362,
      "learning_rate": 6.947819411632222e-09,
      "logits/chosen": -2.229992389678955,
      "logits/rejected": -2.149827241897583,
      "logps/chosen": -371.315185546875,
      "logps/pi_response": -153.27757263183594,
      "logps/ref_response": -143.07461547851562,
      "logps/rejected": -742.3646240234375,
      "loss": 0.4965,
      "rewards/accuracies": 0.762499988079071,
      "rewards/chosen": -0.6061862707138062,
      "rewards/margins": 0.8908622860908508,
      "rewards/rejected": -1.4970484972000122,
      "step": 50
    },
    {
      "epoch": 0.9874476987447699,
      "step": 59,
      "total_flos": 0.0,
      "train_loss": 0.5632201857486013,
      "train_runtime": 2574.0561,
      "train_samples_per_second": 5.937,
      "train_steps_per_second": 0.023
    }
  ],
  "logging_steps": 10,
  "max_steps": 59,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}