{
    "epoch": 1.0,
    "eval_logits/chosen": -2.8154520988464355,
    "eval_logits/rejected": -2.8132472038269043,
    "eval_logps/chosen": -31.453943252563477,
    "eval_logps/rejected": -35.06647872924805,
    "eval_loss": 0.907057523727417,
    "eval_rewards/accuracies": 0.5714285373687744,
    "eval_rewards/chosen": -0.08574554324150085,
    "eval_rewards/margins": 0.0984167754650116,
    "eval_rewards/rejected": -0.18416231870651245,
    "eval_runtime": 112.8077,
    "eval_samples": 343,
    "eval_samples_per_second": 3.041,
    "eval_steps_per_second": 0.381,
    "train_loss": 0.694122297732861,
    "train_runtime": 2717.9333,
    "train_samples": 3079,
    "train_samples_per_second": 1.133,
    "train_steps_per_second": 0.142
}