{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NACBXNL34U36WBVFAYSZYKRVVL","short_pith_number":"pith:NACBXNL3","schema_version":"1.0","canonical_sha256":"68041bb57be537eb06a506259c2a35aac19fa23f5b104cfcd206406c681be2a7","source":{"kind":"arxiv","id":"2403.18715","version":2},"attestation_state":"computed","paper":{"title":"Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.MM"],"primary_cat":"cs.CV","authors_text":"Chris Biemann, Jingheng Pan, Liang Ding, Xintong Wang","submitted_at":"2024-03-27T16:04:47Z","abstract_excerpt":"Large Vision-Language Models (LVLMs) are increasingly adept at generating contextually detailed and coherent responses from visual inputs. However, their application in multimodal decision-making and open-ended generation is hindered by a notable rate of hallucinations, where generated text inaccurately represents the visual contents. To address this issue, this paper introduces the Instruction Contrastive Decoding (ICD) method, a novel approach designed to reduce hallucinations during LVLM inference. Our method is inspired by our observation that what we call disturbance instructions signific"},"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":"2403.18715","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-03-27T16:04:47Z","cross_cats_sorted":["cs.AI","cs.CL","cs.MM"],"title_canon_sha256":"f62fcb60b61136d2fd091960b0a7c4f15ac9921ab79f97e5c6f1e240a2b57090","abstract_canon_sha256":"71a84c1183faaf457590ed95a66219bc8cd3ed9f61cca1263b92c82b837c14c4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:27:47.807712Z","signature_b64":"Xed6Q3YVYQaMiT+rsKgGHMo8Y7+0fXXZfPgpkUmEAuSm6R4Iy83FL1irl2KW6yA3YljaRRx/sxEsoJsUeHGvDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"68041bb57be537eb06a506259c2a35aac19fa23f5b104cfcd206406c681be2a7","last_reissued_at":"2026-07-05T08:27:47.807193Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:27:47.807193Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decoding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.MM"],"primary_cat":"cs.CV","authors_text":"Chris Biemann, Jingheng Pan, Liang Ding, Xintong Wang","submitted_at":"2024-03-27T16:04:47Z","abstract_excerpt":"Large Vision-Language Models (LVLMs) are increasingly adept at generating contextually detailed and coherent responses from visual inputs. However, their application in multimodal decision-making and open-ended generation is hindered by a notable rate of hallucinations, where generated text inaccurately represents the visual contents. To address this issue, this paper introduces the Instruction Contrastive Decoding (ICD) method, a novel approach designed to reduce hallucinations during LVLM inference. Our method is inspired by our observation that what we call disturbance instructions signific"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.18715","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/2403.18715/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":"2403.18715","created_at":"2026-07-05T08:27:47.807259+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.18715v2","created_at":"2026-07-05T08:27:47.807259+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.18715","created_at":"2026-07-05T08:27:47.807259+00:00"},{"alias_kind":"pith_short_12","alias_value":"NACBXNL34U36","created_at":"2026-07-05T08:27:47.807259+00:00"},{"alias_kind":"pith_short_16","alias_value":"NACBXNL34U36WBVF","created_at":"2026-07-05T08:27:47.807259+00:00"},{"alias_kind":"pith_short_8","alias_value":"NACBXNL3","created_at":"2026-07-05T08:27:47.807259+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27596","citing_title":"Dismantling Pathological Shortcuts: A Causal Framework for Faithful LVLM Decoding","ref_index":84,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29805","citing_title":"Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29805","citing_title":"Clearer Sight, Fewer Lies: Oriented Pickup Preference Optimization for Multimodal Hallucination Mitigation","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2406.10185","citing_title":"Detecting and Evaluating Medical Hallucinations in Large Vision Language Models","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2505.06907","citing_title":"A Survey on Foundation Models for Personalized Federated Intelligence","ref_index":220,"is_internal_anchor":false},{"citing_arxiv_id":"2505.13255","citing_title":"Policy Contrastive Decoding for Robotic Foundation Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21919","citing_title":"SDGBiasBench: Benchmarking and Mitigating Vision--Language Models' Biases in Sustainable Development Goals","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2506.09522","citing_title":"Revisit What You See: Revealing Visual Semantics in Vision Tokens to Guide LVLM Decoding","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12384","citing_title":"Scalable Token-Level Hallucination Detection in Large Language Models","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11559","citing_title":"When Looking Is Not Enough: Visual Attention Structure Reveals Hallucination in MLLMs","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25642","citing_title":"Prefill-Time Intervention for Mitigating Hallucination in Large Vision-Language Models","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2404.18930","citing_title":"Hallucination of Multimodal Large Language Models: A Survey","ref_index":170,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09749","citing_title":"See Fair, Speak Truth: Equitable Attention Improves Grounding and Reduces Hallucination in Vision-Language Alignment","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2604.10071","citing_title":"Spotlight and Shadow: Attention-Guided Dual-Anchor Introspective Decoding for MLLM Hallucination Mitigation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17982","citing_title":"Mitigating Multimodal Hallucination via Phase-wise Self-reward","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04641","citing_title":"CAST: Mitigating Object Hallucination in Large Vision-Language Models via Caption-Guided Visual Attention Steering","ref_index":71,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL","json":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL.json","graph_json":"https://pith.science/api/pith-number/NACBXNL34U36WBVFAYSZYKRVVL/graph.json","events_json":"https://pith.science/api/pith-number/NACBXNL34U36WBVFAYSZYKRVVL/events.json","paper":"https://pith.science/paper/NACBXNL3"},"agent_actions":{"view_html":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL","download_json":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL.json","view_paper":"https://pith.science/paper/NACBXNL3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.18715&json=true","fetch_graph":"https://pith.science/api/pith-number/NACBXNL34U36WBVFAYSZYKRVVL/graph.json","fetch_events":"https://pith.science/api/pith-number/NACBXNL34U36WBVFAYSZYKRVVL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL/action/storage_attestation","attest_author":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL/action/author_attestation","sign_citation":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL/action/citation_signature","submit_replication":"https://pith.science/pith/NACBXNL34U36WBVFAYSZYKRVVL/action/replication_record"}},"created_at":"2026-07-05T08:27:47.807259+00:00","updated_at":"2026-07-05T08:27:47.807259+00:00"}