{ "epoch": 3.0, "eval_logits/chosen": -2.2436294555664062, "eval_logits/rejected": -2.2173898220062256, "eval_logps/chosen": -248.40806579589844, "eval_logps/rejected": -218.735107421875, "eval_loss": 0.6023992896080017, "eval_rewards/accuracies": 0.7916666865348816, "eval_rewards/chosen": 1.2815405130386353, "eval_rewards/margins": 2.2554075717926025, "eval_rewards/rejected": -0.9738670587539673, "eval_runtime": 125.4397, "eval_samples_per_second": 15.147, "eval_steps_per_second": 0.239 }