{"slug": "self-distilled-policy-gradient", "title": "Self-Distilled Policy Gradient", "summary": "Researchers introduced SDPG, a self-distilled policy-gradient framework that combines group-relative verifier advantages with normalized standard deviation and full-vocabulary on-policy self-distillation to improve reinforcement learning for language models. The method, which treats self-distillation as an auxiliary reverse Kullback-Leibler divergence loss, demonstrated enhanced stability and performance over existing RLVR and self-distillation baselines in empirical tests.", "body_md": "# Computer Science > Machine Learning\n\n[Submitted on 2 Jun 2026]\n\n# Title:Self-Distilled Policy Gradient\n\n[View PDF](/pdf/2606.04036)\n\n[HTML (experimental)](https://arxiv.org/html/2606.04036v1)\n\nAbstract: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[this https URL].\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer\n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers\n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps\n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations\n\n*(*[What are Smart Citations?](https://www.scite.ai/))# Code, Data and Media Associated with this Article\n\nalphaXiv\n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers\n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub\n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub\n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face\n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast\n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower\n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender\n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\nIArxiv Recommender\n\n*(*[What is IArxiv?](https://iarxiv.org/about))# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.\n\nHave an idea for a project that will add value for arXiv's community? [ Learn more about arXivLabs](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/self-distilled-policy-gradient", "canonical_source": "https://arxiv.org/abs/2606.04036", "published_at": "2026-06-04 04:00:00+00:00", "updated_at": "2026-06-04 04:36:45.980746+00:00", "lang": "en", "topics": ["machine-learning", "large-language-models", "artificial-intelligence", "neural-networks"], "entities": ["SDPG", "RLVR"], "alternates": {"html": "https://wpnews.pro/news/self-distilled-policy-gradient", "markdown": "https://wpnews.pro/news/self-distilled-policy-gradient.md", "text": "https://wpnews.pro/news/self-distilled-policy-gradient.txt", "jsonld": "https://wpnews.pro/news/self-distilled-policy-gradient.jsonld"}}