{"slug": "grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt", "title": "GRP-Obliteration: Unaligning LLMs with a Single Unlabeled Prompt", "summary": "Researchers submitted a paper to arXiv on 5 Feb 2026 introducing GRP-Obliteration (GRP-Oblit), a method that uses Group Relative Policy Optimization (GRPO) to remove safety constraints from aligned models using a single unlabeled prompt. GRP-Oblit achieved stronger unalignment on average than existing state-of-the-art techniques while largely preserving model utility, and generalizes to diffusion-based image generation systems. The method was evaluated on six utility benchmarks and five safety benchmarks across fifteen 7-20B parameter models, including GPT-OSS, distilled DeepSeek, Gemma, Llama, Ministral, and Qwen, spanning instruct and reasoning models and dense and MoE architectures.", "body_md": "# Computer Science > Machine Learning\n\n  [Submitted on 5 Feb 2026]\n\n# Title:GRP-Obliteration: Unaligning LLMs With a Single Unlabeled Prompt\n\n[View PDF](https://arxiv.org/pdf/2602.06258)\n\n[HTML (experimental)](https://arxiv.org/html/2602.06258v1)\n\nAbstract:Safety alignment is only as robust as its weakest failure mode. Despite extensive work on safety post-training, it has been shown that models can be readily unaligned through post-deployment fine-tuning. However, these methods often require extensive data curation and degrade model utility.\n\nIn this work, we extend the practical limits of unalignment by introducing GRP-Obliteration (GRP-Oblit), a method that uses Group Relative Policy Optimization (GRPO) to directly remove safety constraints from target models. We show that a single unlabeled prompt is sufficient to reliably unalign safety-aligned models while largely preserving their utility, and that GRP-Oblit achieves stronger unalignment on average than existing state-of-the-art techniques. Moreover, GRP-Oblit generalizes beyond language models and can also unalign diffusion-based image generation systems.\n\nWe evaluate GRP-Oblit on six utility benchmarks and five safety benchmarks across fifteen 7-20B parameter models, spanning instruct and reasoning models, as well as dense and MoE architectures. The evaluated model families include GPT-OSS, distilled DeepSeek, Gemma, Llama, Ministral, and Qwen.\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/))\n# 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))\n# 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))\n# 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/grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt", "canonical_source": "https://arxiv.org/abs/2602.06258", "published_at": "2026-09-15 14:31:23+00:00", "updated_at": "2026-09-15 14:53:44.209093+00:00", "lang": "en", "topics": ["ai-safety", "large-language-models", "ai-research", "machine-learning", "generative-ai"], "entities": ["arXiv", "GRP-Obliteration", "Group Relative Policy Optimization", "GPT-OSS", "DeepSeek", "Gemma", "Llama", "Qwen"], "alternates": {"html": "https://wpnews.pro/news/grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt", "markdown": "https://wpnews.pro/news/grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt.md", "text": "https://wpnews.pro/news/grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt.txt", "jsonld": "https://wpnews.pro/news/grp-obliteration-unaligning-llms-with-a-single-unlabeled-prompt.jsonld"}}