{"ID":5937077,"CreatedAt":"2026-07-07T03:14:33.014478982Z","UpdatedAt":"2026-07-09T13:28:50.14143324Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.05069","arxiv_id":"2607.05069","title":"MIRAGE: Defending Long-Form RAG Against Misinformation Pollution","abstract":"Retrieval-Augmented Generation (RAG) improves factuality by grounding LLMs in external evidence, but real-world retrieval is often polluted: semantically relevant passages may contain subtle misinformation, misleading framings, or fabrications. We introduce MIRAGE, a training-free, model-agnostic defense for long-form RAG. MIRAGE builds an NLI-based cross-document claim graph and applies a Defended-Claims Gate to either condition generation on a consistent, multi-source supported subset or to block retrieval and answer parametrically. We also release a minimal-edit pollution protocol spanning four perturbation families (Unambiguous, Conflicting, Misleading, Fabricated) to construct matched clean, mixed, and fully polluted evaluation regimes. Across four long-form QA benchmarks and multiple commercial and open-weight LLMs, pollution severely degrades vanilla RAG, while MIRAGE consistently restores factuality under mixed and fully polluted evidence and outperforms prior robust-RAG methods. Our implementation and datasets are available at https://github.com/SaadElDine/MIRAGE.","short_abstract":"Retrieval-Augmented Generation (RAG) improves factuality by grounding LLMs in external evidence, but real-world retrieval is often polluted: semantically relevant passages may contain subtle misinformation, misleading framings, or fabrications. We introduce MIRAGE, a training-free, model-agnostic defense for long-form...","url_abs":"https://arxiv.org/abs/2607.05069","url_pdf":"https://arxiv.org/pdf/2607.05069v1","authors":"[\"Saadeldine Eletter\",\"Ruihong Zeng\",\"Yuxia Wang\",\"Maxim Panov\",\"Aleksandr Rubashevskii\",\"Preslav Nakov\"]","published":"2026-07-06T13:36:12Z","proceeding":"cs.CL","tasks":"[\"cs.CL\"]","methods":"[\"RAG\",\"Large Language Model\"]","has_code":false,"code_links":[{"ID":613956,"CreatedAt":"2026-07-07T03:14:33.014478982Z","UpdatedAt":"2026-07-07T03:14:33.014478982Z","DeletedAt":null,"paper_id":5937077,"paper_url":"https://arxiv.org/abs/2607.05069","paper_title":"MIRAGE: Defending Long-Form RAG Against Misinformation Pollution","repo_url":"https://github.com/SaadElDine/MIRAGE","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}