{
    "epoch": 1.0,
    "eval_logits/chosen": -2.8098936080932617,
    "eval_logits/rejected": -2.807415723800659,
    "eval_logps/chosen": -31.408082962036133,
    "eval_logps/rejected": -34.975181579589844,
    "eval_loss": 0.9007851481437683,
    "eval_rewards/accuracies": 0.5859634280204773,
    "eval_rewards/chosen": -0.1005057543516159,
    "eval_rewards/margins": 0.12111683934926987,
    "eval_rewards/rejected": -0.22162257134914398,
    "eval_runtime": 112.7349,
    "eval_samples": 343,
    "eval_samples_per_second": 3.043,
    "eval_steps_per_second": 0.381
}