{
    "epoch": 3.0,
    "eval_logits/chosen": 1.6027365922927856,
    "eval_logits/rejected": 1.7115638256072998,
    "eval_logps/chosen": -273.4245910644531,
    "eval_logps/rejected": -344.37725830078125,
    "eval_loss": 0.7364417314529419,
    "eval_rewards/accuracies": 0.3194444477558136,
    "eval_rewards/chosen": -2.318794012069702,
    "eval_rewards/margins": 0.7496706247329712,
    "eval_rewards/rejected": -3.068464756011963,
    "eval_runtime": 113.7849,
    "eval_samples": 2000,
    "eval_samples_per_second": 17.577,
    "eval_steps_per_second": 0.554,
    "total_flos": 0.0,
    "train_loss": 0.5020060315728188,
    "train_runtime": 3747.4353,
    "train_samples": 6113,
    "train_samples_per_second": 4.894,
    "train_steps_per_second": 0.077
}