{
    "epoch": 2.999297541394882,
    "eval_logits/chosen": 7.7498650550842285,
    "eval_logits/rejected": 7.4245452880859375,
    "eval_logps/chosen": -45.089073181152344,
    "eval_logps/rejected": -46.10942077636719,
    "eval_loss": 1.940507559083926e-07,
    "eval_rewards/accuracies": 0.5682492852210999,
    "eval_rewards/chosen": -45.089073181152344,
    "eval_rewards/margins": 1.0203527212142944,
    "eval_rewards/rejected": -46.10942077636719,
    "eval_runtime": 39.9113,
    "eval_samples": 1345,
    "eval_samples_per_second": 33.7,
    "eval_steps_per_second": 8.444
}