Probability-Entropy Calibration: An Elastic Indicator for Adaptive Fine-tuning 事件

Probability-Entropy Calibration: An Elastic Indicator for Adaptive Fine-tuning arXiv:2602.01745v2 Announce Type: replace-cross Abstract: Token-level reweighting is a simple yet effective mechanism for controlling supervised fine-tuning, but common indicators are largely one-dimensional: the ground-truth probability reflects downstream alignment, while token entropy reflects intrinsic uncertainty induced by the pre-training prior. Ignoring entropy can misidentify noisy or easily replaceable toke