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

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

来源：arXiv · arxiv.org