#artificial intelligence
Mar 2026
HEAL: Hindsight Entropy-Assisted Learning for Reasoning Distillation
Hindsight Entropy-Assisted Learning (HEAL), an RL-free framework designed to bridge this reasoning gap, is proposed and extensive experiments demonstrate that HEAL significantly outperforms traditional SFT distillation and other baselines.
Wen-Jing Zhang, Jiang-Ze Yan, Jie-Fu Huang et al.
· arXiv.org · 1 citation