{"ID":2886863,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.02296","arxiv_id":"2508.02296","title":"Knowing When Not to Answer: Lightweight KB-Aligned OOD Detection for Safe RAG","abstract":"Retrieval-Augmented Generation (RAG) systems are increasingly deployed in high-stakes domains, where safety depends not only on how a system answers, but also on whether a query should be answered given a knowledge base (KB). Out-of-domain (OOD) queries can cause dense retrieval to surface weakly related context and lead the generator to produce fluent but unjustified responses. We study lightweight, KB-aligned OOD detection as an always-on gate for RAG systems. Our approach applies PCA to KB embeddings and scores queries in a compact subspace selected either by explained-variance retention (EVR) or by a separability-driven t-test ranking. We evaluate geometric semantic-search rules and lightweight classifiers across 16 domains, including high-stakes COVID-19 and Substance Use KBs, and stress-test robustness using both LLM-generated attacks and an in-the-wild 4chan attack. We find that low-dimensional detectors achieve competitive OOD performance while being faster, cheaper, and more interpretable than prompted LLM-based judges. Finally, human and LLM-based evaluations show that OOD queries primarily degrade the relevance of RAG outputs, showing the need for efficient external OOD detection to maintain safe, in-scope behavior.","short_abstract":"Retrieval-Augmented Generation (RAG) systems are increasingly deployed in high-stakes domains, where safety depends not only on how a system answers, but also on whether a query should be answered given a knowledge base (KB). Out-of-domain (OOD) queries can cause dense retrieval to surface weakly related context and le...","url_abs":"https://arxiv.org/abs/2508.02296","url_pdf":"https://arxiv.org/pdf/2508.02296v2","authors":"[\"Ilias Triantafyllopoulos\",\"Renyi Qu\",\"Salvatore Giorgi\",\"Brenda Curtis\",\"Lyle H. Ungar\",\"João Sedoc\"]","published":"2025-08-04T11:04:54Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.IR\"]","methods":"[\"RAG\",\"Large Language Model\"]","has_code":false}