{
    "epoch": 2.9969690846635686,
    "eval_logits/chosen": -2.8472585678100586,
    "eval_logits/rejected": -2.8558220863342285,
    "eval_logps/chosen": -0.7975095510482788,
    "eval_logps/rejected": -1.0328320264816284,
    "eval_loss": 0.8629826903343201,
    "eval_odds_ratio_loss": 0.6547309160232544,
    "eval_rewards/accuracies": 0.5618181824684143,
    "eval_rewards/chosen": -0.07975095510482788,
    "eval_rewards/margins": 0.02353225089609623,
    "eval_rewards/rejected": -0.10328320413827896,
    "eval_runtime": 194.6972,
    "eval_samples_per_second": 5.65,
    "eval_sft_loss": 0.7975095510482788,
    "eval_steps_per_second": 2.825,
    "total_flos": 2.1013894560546816e+18,
    "train_loss": 0.9013287582572352,
    "train_runtime": 18144.1457,
    "train_samples_per_second": 1.637,
    "train_steps_per_second": 0.102
}