{"ID":2874786,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.04655","arxiv_id":"2509.04655","title":"Polysemantic Dropout: Conformal OOD Detection for Specialized LLMs","abstract":"We propose a novel inference-time out-of-domain (OOD) detection algorithm for specialized large language models (LLMs). Despite achieving state-of-the-art performance on in-domain tasks through fine-tuning, specialized LLMs remain vulnerable to incorrect or unreliable outputs when presented with OOD inputs, posing risks in critical applications. Our method leverages the Inductive Conformal Anomaly Detection (ICAD) framework, using a new non-conformity measure based on the model's dropout tolerance. Motivated by recent findings on polysemanticity and redundancy in LLMs, we hypothesize that in-domain inputs exhibit higher dropout tolerance than OOD inputs. We aggregate dropout tolerance across multiple layers via a valid ensemble approach, improving detection while maintaining theoretical false alarm bounds from ICAD. Experiments with medical-specialized LLMs show that our approach detects OOD inputs better than baseline methods, with AUROC improvements of $2\\%$ to $37\\%$ when treating OOD datapoints as positives and in-domain test datapoints as negatives.","short_abstract":"We propose a novel inference-time out-of-domain (OOD) detection algorithm for specialized large language models (LLMs). Despite achieving state-of-the-art performance on in-domain tasks through fine-tuning, specialized LLMs remain vulnerable to incorrect or unreliable outputs when presented with OOD inputs, posing risk...","url_abs":"https://arxiv.org/abs/2509.04655","url_pdf":"https://arxiv.org/pdf/2509.04655v2","authors":"[\"Ayush Gupta\",\"Ramneet Kaur\",\"Anirban Roy\",\"Adam D. Cobb\",\"Rama Chellappa\",\"Susmit Jha\"]","published":"2025-09-04T20:50:51Z","proceeding":"cs.CL","tasks":"[\"cs.CL\",\"cs.AI\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}