Prune-OPD: Efficient and Reliable On-Policy Distillation for Long-Horizon Reasoning

概要

arXiv:2605.07804v1 Announce Type: cross Abstract: On-policy distillation (OPD) leverages dense teacher rewards to enhance reasoning models. However, scaling OPD to long-horizon tasks exposes a critical flaw: as the student's generated prefix inevitably diverges from the teacher's thought process, t…