{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5CTSE3WST6I6DAP4H676JBEBXT","short_pith_number":"pith:5CTSE3WS","schema_version":"1.0","canonical_sha256":"e8a7226ed29f91e181fc3fbfe48481bcdef99ebe328fd70fbd312c2bec4d4f84","source":{"kind":"arxiv","id":"2505.08200","version":1},"attestation_state":"computed","paper":{"title":"A Head to Predict and a Head to Question: Pre-trained Uncertainty Quantification Heads for Hallucination Detection in LLM Outputs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Akim Tsvigun, Artem Shelmanov, Artem Vazhentsev, Caiqi Zhang, Ekaterina Fadeeva, Igor Kiselev, Ivan Tsvigun, Mrinmaya Sachan, Nico Daheim, Preslav Nakov, Timothy Baldwin, Zhuohan Xie","submitted_at":"2025-05-13T03:30:26Z","abstract_excerpt":"Large Language Models (LLMs) have the tendency to hallucinate, i.e., to sporadically generate false or fabricated information. This presents a major challenge, as hallucinations often appear highly convincing and users generally lack the tools to detect them. Uncertainty quantification (UQ) provides a framework for assessing the reliability of model outputs, aiding in the identification of potential hallucinations. In this work, we introduce pre-trained UQ heads: supervised auxiliary modules for LLMs that substantially enhance their ability to capture uncertainty compared to unsupervised UQ me"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2505.08200","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-13T03:30:26Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f4a647ec692848eb57b756f619a738b56524e15be25819bfdf7a1b9febe8d2a7","abstract_canon_sha256":"c08c96ee020c046b97f314868e07fc3ea70dbcef05afd18d96c97ab97cdd3a33"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:02:26.388775Z","signature_b64":"5r0A7sm9FafCP+ER6rZIE83gOIyvGrO69yAoRVQ0eCMSNXSZLNTHiL6bfeM0RTRsWPcp5W9f5jCiR3Kt9DapCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e8a7226ed29f91e181fc3fbfe48481bcdef99ebe328fd70fbd312c2bec4d4f84","last_reissued_at":"2026-07-05T11:02:26.388328Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:02:26.388328Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Head to Predict and a Head to Question: Pre-trained Uncertainty Quantification Heads for Hallucination Detection in LLM Outputs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Akim Tsvigun, Artem Shelmanov, Artem Vazhentsev, Caiqi Zhang, Ekaterina Fadeeva, Igor Kiselev, Ivan Tsvigun, Mrinmaya Sachan, Nico Daheim, Preslav Nakov, Timothy Baldwin, Zhuohan Xie","submitted_at":"2025-05-13T03:30:26Z","abstract_excerpt":"Large Language Models (LLMs) have the tendency to hallucinate, i.e., to sporadically generate false or fabricated information. This presents a major challenge, as hallucinations often appear highly convincing and users generally lack the tools to detect them. Uncertainty quantification (UQ) provides a framework for assessing the reliability of model outputs, aiding in the identification of potential hallucinations. In this work, we introduce pre-trained UQ heads: supervised auxiliary modules for LLMs that substantially enhance their ability to capture uncertainty compared to unsupervised UQ me"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.08200","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2505.08200/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2505.08200","created_at":"2026-07-05T11:02:26.388395+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.08200v1","created_at":"2026-07-05T11:02:26.388395+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.08200","created_at":"2026-07-05T11:02:26.388395+00:00"},{"alias_kind":"pith_short_12","alias_value":"5CTSE3WST6I6","created_at":"2026-07-05T11:02:26.388395+00:00"},{"alias_kind":"pith_short_16","alias_value":"5CTSE3WST6I6DAP4","created_at":"2026-07-05T11:02:26.388395+00:00"},{"alias_kind":"pith_short_8","alias_value":"5CTSE3WS","created_at":"2026-07-05T11:02:26.388395+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT","json":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT.json","graph_json":"https://pith.science/api/pith-number/5CTSE3WST6I6DAP4H676JBEBXT/graph.json","events_json":"https://pith.science/api/pith-number/5CTSE3WST6I6DAP4H676JBEBXT/events.json","paper":"https://pith.science/paper/5CTSE3WS"},"agent_actions":{"view_html":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT","download_json":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT.json","view_paper":"https://pith.science/paper/5CTSE3WS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.08200&json=true","fetch_graph":"https://pith.science/api/pith-number/5CTSE3WST6I6DAP4H676JBEBXT/graph.json","fetch_events":"https://pith.science/api/pith-number/5CTSE3WST6I6DAP4H676JBEBXT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT/action/storage_attestation","attest_author":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT/action/author_attestation","sign_citation":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT/action/citation_signature","submit_replication":"https://pith.science/pith/5CTSE3WST6I6DAP4H676JBEBXT/action/replication_record"}},"created_at":"2026-07-05T11:02:26.388395+00:00","updated_at":"2026-07-05T11:02:26.388395+00:00"}