{
    "epoch": 3.0,
    "eval_logits/chosen": -2.7101802825927734,
    "eval_logits/rejected": -2.6752188205718994,
    "eval_logps/chosen": -321.78399658203125,
    "eval_logps/rejected": -318.9219665527344,
    "eval_loss": 0.6641141176223755,
    "eval_rewards/accuracies": 0.7739999890327454,
    "eval_rewards/chosen": -4.202523708343506,
    "eval_rewards/margins": 3.3038268089294434,
    "eval_rewards/rejected": -7.506350994110107,
    "eval_runtime": 353.2957,
    "eval_samples": 2000,
    "eval_samples_per_second": 5.661,
    "eval_steps_per_second": 0.354
}