{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GMVDS7E5G55M6J7LUY4IPVBCU3","short_pith_number":"pith:GMVDS7E5","schema_version":"1.0","canonical_sha256":"332a397c9d377acf27eba63887d422a6e01817da15e6b057a35dc15c87baf0ca","source":{"kind":"arxiv","id":"2411.11919","version":2},"attestation_state":"computed","paper":{"title":"VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hu Zhang, Ruiyang Zhang, Zhedong Zheng","submitted_at":"2024-11-18T04:06:04Z","abstract_excerpt":"Given the higher information load processed by large vision-language models (LVLMs) compared to single-modal LLMs, detecting LVLM hallucinations requires more human and time expense, and thus rise a wider safety concerns. In this paper, we introduce VL-Uncertainty, the first uncertainty-based framework for detecting hallucinations in LVLMs. Different from most existing methods that require ground-truth or pseudo annotations, VL-Uncertainty utilizes uncertainty as an intrinsic metric. We measure uncertainty by analyzing the prediction variance across semantically equivalent but perturbed prompt"},"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":"2411.11919","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-11-18T04:06:04Z","cross_cats_sorted":[],"title_canon_sha256":"ab3ce53dacb16f8361c7619c0e78f8cf3c353b5e654d86467c2f12b43653b3b6","abstract_canon_sha256":"ccad70498b072befa1cfea7c1da708e706fe49f1bb4f914fc52ed44d7cea7939"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:42:32.897170Z","signature_b64":"EscBUt9ohAr+tjYErskUk96lJ10uzXcVkO3tv5k9CPnXwABqDhfEobZt/dbEt4u3Pgeqqjc+4MgOE2qHRLH5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"332a397c9d377acf27eba63887d422a6e01817da15e6b057a35dc15c87baf0ca","last_reissued_at":"2026-07-05T09:42:32.896681Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:42:32.896681Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hu Zhang, Ruiyang Zhang, Zhedong Zheng","submitted_at":"2024-11-18T04:06:04Z","abstract_excerpt":"Given the higher information load processed by large vision-language models (LVLMs) compared to single-modal LLMs, detecting LVLM hallucinations requires more human and time expense, and thus rise a wider safety concerns. In this paper, we introduce VL-Uncertainty, the first uncertainty-based framework for detecting hallucinations in LVLMs. Different from most existing methods that require ground-truth or pseudo annotations, VL-Uncertainty utilizes uncertainty as an intrinsic metric. We measure uncertainty by analyzing the prediction variance across semantically equivalent but perturbed prompt"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.11919","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/2411.11919/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":"2411.11919","created_at":"2026-07-05T09:42:32.896747+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.11919v2","created_at":"2026-07-05T09:42:32.896747+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.11919","created_at":"2026-07-05T09:42:32.896747+00:00"},{"alias_kind":"pith_short_12","alias_value":"GMVDS7E5G55M","created_at":"2026-07-05T09:42:32.896747+00:00"},{"alias_kind":"pith_short_16","alias_value":"GMVDS7E5G55M6J7L","created_at":"2026-07-05T09:42:32.896747+00:00"},{"alias_kind":"pith_short_8","alias_value":"GMVDS7E5","created_at":"2026-07-05T09:42:32.896747+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25760","citing_title":"Uncertainty Quantification for Computer-Use Agents: A Benchmark across Vision-Language Models and GUI Grounding Datasets","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26923","citing_title":"GAVEL: Grounded Caption Error Verification and Localization","ref_index":100,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20754","citing_title":"Perturbation-Based Uncertainty for Failure Detection in Vision-Language-Action Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31407","citing_title":"Visual Semantic Entropy: Do Vision Language Models Recognize Visual Ambiguity?","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.32012","citing_title":"CoMet: Context and Multiplicity Decomposition for Multimodal Uncertainty Estimation","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20772","citing_title":"VIHD: Visual Intervention-based Hallucination Detection for Medical Visual Question Answering","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30834","citing_title":"Hide-and-Seek in Trajectories: Discovering Failure Signals for VLA Runtime Monitoring","ref_index":82,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20772","citing_title":"VIHD: Visual Intervention-based Hallucination Detection for Medical Visual Question Answering","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20084","citing_title":"BalanceRAG: Joint Risk Calibration for Cascaded Retrieval-Augmented Generation","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10893","citing_title":"Grounded or Guessing? LVLM Confidence Estimation via Blind-Image Contrastive Ranking","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2511.01458","citing_title":"When to Trust the Answer: Question-Aligned Semantic Nearest Neighbor Entropy for Safer Surgical VQA","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10893","citing_title":"Grounded or Guessing? LVLM Confidence Estimation via Blind-Image Contrastive Ranking","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00326","citing_title":"Prompt-Induced Score Variance in Zero-Shot Binary Vision-Language Safety Classification","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12035","citing_title":"When Does Visual Token Pruning Improve Calibration? The Role of Evidence Coverage in MLLMs","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3","json":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3.json","graph_json":"https://pith.science/api/pith-number/GMVDS7E5G55M6J7LUY4IPVBCU3/graph.json","events_json":"https://pith.science/api/pith-number/GMVDS7E5G55M6J7LUY4IPVBCU3/events.json","paper":"https://pith.science/paper/GMVDS7E5"},"agent_actions":{"view_html":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3","download_json":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3.json","view_paper":"https://pith.science/paper/GMVDS7E5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.11919&json=true","fetch_graph":"https://pith.science/api/pith-number/GMVDS7E5G55M6J7LUY4IPVBCU3/graph.json","fetch_events":"https://pith.science/api/pith-number/GMVDS7E5G55M6J7LUY4IPVBCU3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3/action/storage_attestation","attest_author":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3/action/author_attestation","sign_citation":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3/action/citation_signature","submit_replication":"https://pith.science/pith/GMVDS7E5G55M6J7LUY4IPVBCU3/action/replication_record"}},"created_at":"2026-07-05T09:42:32.896747+00:00","updated_at":"2026-07-05T09:42:32.896747+00:00"}