{
    "epoch": 1.9905213270142181,
    "eval_logits/chosen": -1.3085236549377441,
    "eval_logits/rejected": -1.310121774673462,
    "eval_logps/chosen": -325.1966552734375,
    "eval_logps/rejected": -340.3136291503906,
    "eval_loss": 0.4830668568611145,
    "eval_rewards/accuracies": 0.7708333134651184,
    "eval_rewards/chosen": -0.2542985677719116,
    "eval_rewards/margins": 0.8397493958473206,
    "eval_rewards/rejected": -1.094048023223877,
    "eval_runtime": 37.7518,
    "eval_samples": 750,
    "eval_samples_per_second": 19.867,
    "eval_steps_per_second": 0.636,
    "total_flos": 0.0,
    "train_loss": 0.5004417010716029,
    "train_runtime": 1390.6464,
    "train_samples": 6750,
    "train_samples_per_second": 9.708,
    "train_steps_per_second": 0.151
}