AI RESEARCH
ROSD: Reflective On-Policy Self-Distillation for Language Model Reasoning across Domains
arXiv CS.LG
•
ArXi:2605.28014v1 Announce Type: cross On-policy self-distillation (OPSD) improves the reasoning performance of large language models (LLMs) by providing dense token-level supervision for on-policy rollouts. However, existing OPSD methods often yield limited gains on in-domain reasoning and generalize poorly to out-of-domain problems. We identify two key causes: conditioning the self-teacher on a verified solution encourages imitation of