{
    "epoch": 3.0,
    "eval_logits/chosen": -2.5570387840270996,
    "eval_logits/rejected": -2.5411317348480225,
    "eval_logps/chosen": -74.19338989257812,
    "eval_logps/rejected": -82.50736999511719,
    "eval_loss": 0.9815396666526794,
    "eval_rewards/accuracies": 0.3452380895614624,
    "eval_rewards/chosen": 0.2789943516254425,
    "eval_rewards/margins": 1.5707836151123047,
    "eval_rewards/rejected": -1.2917892932891846,
    "eval_runtime": 113.7681,
    "eval_samples": 2000,
    "eval_samples_per_second": 17.58,
    "eval_steps_per_second": 0.554,
    "total_flos": 0.0,
    "train_loss": 0.479405563738611,
    "train_runtime": 3137.8924,
    "train_samples": 6113,
    "train_samples_per_second": 5.844,
    "train_steps_per_second": 0.092
}