{
    "epoch": 0.9987943737441393,
    "eval_logits/chosen": -3.4656760692596436,
    "eval_logits/rejected": -3.4604122638702393,
    "eval_logps/chosen": -0.5584612488746643,
    "eval_logps/rejected": -0.6149689555168152,
    "eval_loss": 1.5217351913452148,
    "eval_rewards/accuracies": 0.5811170339584351,
    "eval_rewards/chosen": -0.5584612488746643,
    "eval_rewards/margins": 0.05650780722498894,
    "eval_rewards/rejected": -0.6149689555168152,
    "eval_runtime": 435.8552,
    "eval_samples": 2994,
    "eval_samples_per_second": 6.869,
    "eval_steps_per_second": 0.431
}