{
    "epoch": 1.0,
    "eval_logits/chosen": NaN,
    "eval_logits/rejected": 1.1838418245315552,
    "eval_logps/chosen": -350.3968200683594,
    "eval_logps/rejected": -322.0793762207031,
    "eval_loss": NaN,
    "eval_nll_loss": NaN,
    "eval_rewards/accuracies": 0.6478174328804016,
    "eval_rewards/chosen": 1.7388392686843872,
    "eval_rewards/margins": 0.6021825671195984,
    "eval_rewards/rejected": 1.136191725730896,
    "eval_runtime": 8.4094,
    "eval_samples_per_second": 118.914,
    "eval_steps_per_second": 7.492
}