{
    "epoch": 1.0,
    "eval_logits/chosen": -2.850416660308838,
    "eval_logits/rejected": -2.823277235031128,
    "eval_logps/chosen": -280.9537353515625,
    "eval_logps/rejected": -262.90020751953125,
    "eval_loss": 0.6776002049446106,
    "eval_rewards/accuracies": 0.6855000257492065,
    "eval_rewards/chosen": 0.018197493627667427,
    "eval_rewards/margins": 0.032815974205732346,
    "eval_rewards/rejected": -0.014618484303355217,
    "eval_runtime": 623.229,
    "eval_samples": 2000,
    "eval_samples_per_second": 3.209,
    "eval_steps_per_second": 0.401,
    "total_flos": 0.0,
    "train_loss": 0.680465580483626,
    "train_runtime": 64957.9706,
    "train_samples": 61134,
    "train_samples_per_second": 0.941,
    "train_steps_per_second": 0.059
}