{
    "epoch": 1.971563981042654,
    "eval_logits/chosen": 96.71578216552734,
    "eval_logits/rejected": 90.98221588134766,
    "eval_logps/chosen": -423.6227722167969,
    "eval_logps/rejected": -453.7782287597656,
    "eval_loss": 0.468290776014328,
    "eval_rewards/accuracies": 0.7708333134651184,
    "eval_rewards/chosen": -3.0221338272094727,
    "eval_rewards/margins": 1.6591955423355103,
    "eval_rewards/rejected": -4.681329727172852,
    "eval_runtime": 58.6185,
    "eval_samples": 750,
    "eval_samples_per_second": 12.795,
    "eval_steps_per_second": 0.409,
    "total_flos": 0.0,
    "train_loss": 0.3921648321243433,
    "train_runtime": 1190.3032,
    "train_samples": 6750,
    "train_samples_per_second": 11.342,
    "train_steps_per_second": 0.087
}