{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:E6UIPIDKTIQZAP3GTT3UW35525","short_pith_number":"pith:E6UIPIDK","schema_version":"1.0","canonical_sha256":"27a887a06a9a21903f669cf74b6fbdd754d24c320a6b0939871d745bd994716d","source":{"kind":"arxiv","id":"2502.13622","version":2},"attestation_state":"computed","paper":{"title":"REFIND at SemEval-2025 Task 3: Retrieval-Augmented Factuality Hallucination Detection in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"DongGeon Lee, HwanJo Yu","submitted_at":"2025-02-19T10:59:05Z","abstract_excerpt":"Hallucinations in large language model (LLM) outputs severely limit their reliability in knowledge-intensive tasks such as question answering. To address this challenge, we introduce REFIND (Retrieval-augmented Factuality hallucINation Detection), a novel framework that detects hallucinated spans within LLM outputs by directly leveraging retrieved documents. As part of the REFIND, we propose the Context Sensitivity Ratio (CSR), a novel metric that quantifies the sensitivity of LLM outputs to retrieved evidence. This innovative approach enables REFIND to efficiently and accurately detect halluc"},"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":"2502.13622","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-19T10:59:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"339e6e56c89952857a20ad482df48c9be6b757a0cf324185d1c08555dcbeadd3","abstract_canon_sha256":"e94e487ee07b422a10297480e3aaee688df665bb71a0e83716aafd1d61841733"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:46:00.506283Z","signature_b64":"dvq7F1Iiq4Q7lUZtC5Q2uBx6ru1jpjJPz5OVhPlHl3CXGNElg0LtTwF8FTswsiBxZS+R2tyYYdzA9QWGJ45dCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"27a887a06a9a21903f669cf74b6fbdd754d24c320a6b0939871d745bd994716d","last_reissued_at":"2026-07-05T10:46:00.505823Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:46:00.505823Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REFIND at SemEval-2025 Task 3: Retrieval-Augmented Factuality Hallucination Detection in Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"DongGeon Lee, HwanJo Yu","submitted_at":"2025-02-19T10:59:05Z","abstract_excerpt":"Hallucinations in large language model (LLM) outputs severely limit their reliability in knowledge-intensive tasks such as question answering. To address this challenge, we introduce REFIND (Retrieval-augmented Factuality hallucINation Detection), a novel framework that detects hallucinated spans within LLM outputs by directly leveraging retrieved documents. As part of the REFIND, we propose the Context Sensitivity Ratio (CSR), a novel metric that quantifies the sensitivity of LLM outputs to retrieved evidence. This innovative approach enables REFIND to efficiently and accurately detect halluc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13622","kind":"arxiv","version":2},"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/2502.13622/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":"2502.13622","created_at":"2026-07-05T10:46:00.505880+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13622v2","created_at":"2026-07-05T10:46:00.505880+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13622","created_at":"2026-07-05T10:46:00.505880+00:00"},{"alias_kind":"pith_short_12","alias_value":"E6UIPIDKTIQZ","created_at":"2026-07-05T10:46:00.505880+00:00"},{"alias_kind":"pith_short_16","alias_value":"E6UIPIDKTIQZAP3G","created_at":"2026-07-05T10:46:00.505880+00:00"},{"alias_kind":"pith_short_8","alias_value":"E6UIPIDK","created_at":"2026-07-05T10:46:00.505880+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.05244","citing_title":"Towards Dependable Retrieval-Augmented Generation Using Factual Confidence Prediction","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525","json":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525.json","graph_json":"https://pith.science/api/pith-number/E6UIPIDKTIQZAP3GTT3UW35525/graph.json","events_json":"https://pith.science/api/pith-number/E6UIPIDKTIQZAP3GTT3UW35525/events.json","paper":"https://pith.science/paper/E6UIPIDK"},"agent_actions":{"view_html":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525","download_json":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525.json","view_paper":"https://pith.science/paper/E6UIPIDK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13622&json=true","fetch_graph":"https://pith.science/api/pith-number/E6UIPIDKTIQZAP3GTT3UW35525/graph.json","fetch_events":"https://pith.science/api/pith-number/E6UIPIDKTIQZAP3GTT3UW35525/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525/action/storage_attestation","attest_author":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525/action/author_attestation","sign_citation":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525/action/citation_signature","submit_replication":"https://pith.science/pith/E6UIPIDKTIQZAP3GTT3UW35525/action/replication_record"}},"created_at":"2026-07-05T10:46:00.505880+00:00","updated_at":"2026-07-05T10:46:00.505880+00:00"}