{
    "epoch": 1.0,
    "eval_logits/chosen": -2.821664571762085,
    "eval_logits/rejected": -2.819110155105591,
    "eval_logps/chosen": -31.076623916625977,
    "eval_logps/rejected": -34.71748352050781,
    "eval_loss": 0.49079883098602295,
    "eval_rewards/accuracies": 0.5867940187454224,
    "eval_rewards/chosen": 0.04116562753915787,
    "eval_rewards/margins": 0.045030828565359116,
    "eval_rewards/rejected": -0.003865197068080306,
    "eval_runtime": 112.8844,
    "eval_samples": 343,
    "eval_samples_per_second": 3.039,
    "eval_steps_per_second": 0.381
}