Weaken to Strengthen: Weakening-Induced On-Policy Self-Distillation for Large Reasoning Models
arXiv:2610.38681v1 Announce Type: new Abstract: Self-distillation offers a way to improve reasoning models when a suitable external teacher is unavailable. However, common approaches rely on privileged information to construct a stronger self-teacher, limiting their use when only unlabeled prompts are available. In this paper, we introduce Weakening-Induced On-Policy Self-Distillation (WISD), which instead constructs a weaker counterpart of the original model and uses the contrast between the t…