{
    "epoch": 5.992211163998269,
    "eval_logits/chosen": -0.271484375,
    "eval_logits/rejected": -0.1357421875,
    "eval_logps/chosen": -110.5,
    "eval_logps/rejected": -129.0,
    "eval_loss": 0.6188219785690308,
    "eval_rewards/accuracies": 0.5,
    "eval_rewards/chosen": 1.0859375,
    "eval_rewards/margins": 4.0,
    "eval_rewards/rejected": -2.90625,
    "eval_runtime": 53.3159,
    "eval_samples_per_second": 72.999,
    "eval_steps_per_second": 2.288
}