{
    "epoch": 0.9987943737441393,
    "eval_logits/chosen": -2.756258010864258,
    "eval_logits/rejected": -2.754289388656616,
    "eval_logps/chosen": -0.6834676861763,
    "eval_logps/rejected": -0.7465164661407471,
    "eval_loss": 0.7904670238494873,
    "eval_rewards/accuracies": 0.605053186416626,
    "eval_rewards/chosen": -1.7086691856384277,
    "eval_rewards/margins": 0.15762200951576233,
    "eval_rewards/rejected": -1.8662911653518677,
    "eval_runtime": 102.313,
    "eval_samples": 2994,
    "eval_samples_per_second": 29.263,
    "eval_steps_per_second": 1.837
}