Weaken to Strengthen: Weakening-Induced On-Policy Self-Distillation for Large Reasoning Models

By Yecheng Wu, Song Han, Han Cai

· HF Daily Papers · Oct 28, 2026