MMODELYST
Papers/Self-Distilled Policy Gradient
PAP

Self-Distilled Policy Gradient

Jun 2, 2026

arXiv
Abstract

On-policy self-distillation, where a language model conditions on privileged context to supervise its own generations, is a promising source of dense supervision for sparse-reward reinforcement learning. Actually, it can be instantiated as an auxiliary full-vocabulary student-to-teacher reverse Kullback-Leibler divergence loss. We therefore propose SDPG, a self-distilled policy-gradient framework that combines group-relative verifier advantages with normalized standard deviation, exact full-vocabulary on-policy self-distillation, as well as reference-policy KL regularization. Empirically, SDPG improves stability and performance over RLVR and self-distillation baselines. The code is available at https://github.com/lauyikfung/SDPG.

Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Yifeng Liu, Shiyuan Zhang, Yifan Zhang, Quanquan Gu
Your notes (browser-local)
saved
Cross-links
No linked entities.
arXiv:2606.04036