{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:JHTW7AMALJK4JSDR3OR2WJXT3F","short_pith_number":"pith:JHTW7AMA","schema_version":"1.0","canonical_sha256":"49e76f81805a55c4c871dba3ab26f3d953edb4d0e81b5a42139b8ca28b71f008","source":{"kind":"arxiv","id":"2402.11411","version":1},"attestation_state":"computed","paper":{"title":"Aligning Modalities in Vision Large Language Models via Preference Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Preference fine-tuning on automatically generated hallucinated responses aligns vision and language modalities in large models while cutting hallucinations.","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chelsea Finn, Chenhang Cui, Huaxiu Yao, Rafael Rafailov, Yiyang Zhou","submitted_at":"2024-02-18T00:56:16Z","abstract_excerpt":"Instruction-following Vision Large Language Models (VLLMs) have achieved significant progress recently on a variety of tasks. These approaches merge strong pre-trained vision models and large language models (LLMs). Since these components are trained separately, the learned representations need to be aligned with joint training on additional image-language pairs. This procedure is not perfect and can cause the model to hallucinate - provide answers that do not accurately reflect the image, even when the core LLM is highly factual and the vision backbone has sufficiently complete representation"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2402.11411","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-18T00:56:16Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"95495881af9bfc34909c83f89fd32e4e9c83ca55cee19e4cb282e8de90b75049","abstract_canon_sha256":"684c4ab3db146800bb4720a51e3c130124f5c1e1b7bb92be651fce3d29e8a9af"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:14.301978Z","signature_b64":"lJwLq7cAoEFNyhF3BP/8U+zA4IL/j5dUzks3bapVbt7RM+FZvbpq+GFeWV09NgW7E8z3wkdgTqjmamnFckLOBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"49e76f81805a55c4c871dba3ab26f3d953edb4d0e81b5a42139b8ca28b71f008","last_reissued_at":"2026-05-17T23:38:14.301452Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:14.301452Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Aligning Modalities in Vision Large Language Models via Preference Fine-tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Preference fine-tuning on automatically generated hallucinated responses aligns vision and language modalities in large models while cutting hallucinations.","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chelsea Finn, Chenhang Cui, Huaxiu Yao, Rafael Rafailov, Yiyang Zhou","submitted_at":"2024-02-18T00:56:16Z","abstract_excerpt":"Instruction-following Vision Large Language Models (VLLMs) have achieved significant progress recently on a variety of tasks. These approaches merge strong pre-trained vision models and large language models (LLMs). Since these components are trained separately, the learned representations need to be aligned with joint training on additional image-language pairs. This procedure is not perfect and can cause the model to hallucinate - provide answers that do not accurately reflect the image, even when the core LLM is highly factual and the vision backbone has sufficiently complete representation"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"In experiments across broad benchmarks, we show that we can not only reduce hallucinations, but improve model performance across standard benchmarks, outperforming prior approaches.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The two-stage automated generation of dispreferred responses (GPT-4V hallucination injection and image distortion) produces high-quality preference pairs that accurately reflect and correct the model's hallucination behavior when used in DPO.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"POVID generates AI-created preference data to fine-tune vision-language models with DPO, reducing hallucinations and improving benchmark scores.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Preference fine-tuning on automatically generated hallucinated responses aligns vision and language modalities in large models while cutting hallucinations.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fc41d789a43373108e2b1e47bfe0168f29991977cfe7ec321abbbf203a40db8a"},"source":{"id":"2402.11411","kind":"arxiv","version":1},"verdict":{"id":"c94e6e5a-3650-40e1-8bb0-efed02110008","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-17T10:54:31.659670Z","strongest_claim":"In experiments across broad benchmarks, we show that we can not only reduce hallucinations, but improve model performance across standard benchmarks, outperforming prior approaches.","one_line_summary":"POVID generates AI-created preference data to fine-tune vision-language models with DPO, reducing hallucinations and improving benchmark scores.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The two-stage automated generation of dispreferred responses (GPT-4V hallucination injection and image distortion) produces high-quality preference pairs that accurately reflect and correct the model's hallucination behavior when used in DPO.","pith_extraction_headline":"Preference fine-tuning on automatically generated hallucinated responses aligns vision and language modalities in large models while cutting hallucinations."},"references":{"count":155,"sample":[{"doi":"","year":2023,"title":"InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning , author=. 2023 , eprint=","work_id":"2a233147-17ae-4a84-a51a-aeaa6303e8e1","ref_index":18,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2000,"title":"Langley , title =","work_id":"6cd283dc-0548-45e9-af07-6bc1005593ad","ref_index":20,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1980,"title":"T. M. Mitchell. The Need for Biases in Learning Generalizations. 1980","work_id":"6b09bca6-ef5d-4a83-8c8e-219f23cbd761","ref_index":21,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"M. J. Kearns , title =","work_id":"8efd8073-6f5d-45c5-94d5-62d366b52518","ref_index":22,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1983,"title":"Machine Learning: An Artificial Intelligence Approach, Vol. I. 1983","work_id":"51835800-f16e-4534-8339-d3ea09147556","ref_index":23,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":155,"snapshot_sha256":"5b1f65218dc218aa0a2bbc2f22c81feb7663e2379061bb550038898361d7e30c","internal_anchors":22},"formal_canon":{"evidence_count":2,"snapshot_sha256":"41ded6d331100e664479d2b235b3d8c2d8b5b1aa5b27fbfa16f9871c6754a422"},"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":"2402.11411","created_at":"2026-05-17T23:38:14.301535+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.11411v1","created_at":"2026-05-17T23:38:14.301535+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.11411","created_at":"2026-05-17T23:38:14.301535+00:00"},{"alias_kind":"pith_short_12","alias_value":"JHTW7AMALJK4","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"JHTW7AMALJK4JSDR","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"JHTW7AMA","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":19,"sample":[{"citing_arxiv_id":"2605.15300","citing_title":"Deep Pre-Alignment for VLMs","ref_index":78,"is_internal_anchor":true},{"citing_arxiv_id":"2407.03320","citing_title":"InternLM-XComposer-2.5: A Versatile Large Vision Language Model Supporting Long-Contextual Input and Output","ref_index":183,"is_internal_anchor":true},{"citing_arxiv_id":"2411.10442","citing_title":"Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization","ref_index":120,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14054","citing_title":"Bad Seeing or Bad Thinking? Rewarding Perception for Vision-Language Reasoning","ref_index":14,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11405","citing_title":"20/20 Vision Language Models: A Prescription for Better VLMs through Data Curation Alone","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2503.01785","citing_title":"Visual-RFT: Visual Reinforcement Fine-Tuning","ref_index":51,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11405","citing_title":"20/20 Vision Language Models: A Prescription for Better VLMs through Data Curation Alone","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.03426","citing_title":"Replacing Parameters with Preferences: Federated Alignment of Heterogeneous Vision-Language Models","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2605.10622","citing_title":"Vocabulary Hijacking in LVLMs: Unveiling Critical Attention Heads by Excluding Inert Tokens to Mitigate Hallucination","ref_index":58,"is_internal_anchor":true},{"citing_arxiv_id":"2605.04874","citing_title":"Uncertainty-Aware Exploratory Direct Preference Optimization for Multimodal Large Language Models","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2605.00323","citing_title":"Online Self-Calibration Against Hallucination in Vision-Language Models","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2604.20696","citing_title":"R-CoV: Region-Aware Chain-of-Verification for Alleviating Object Hallucinations in LVLMs","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19544","citing_title":"DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling","ref_index":66,"is_internal_anchor":true},{"citing_arxiv_id":"2404.18930","citing_title":"Hallucination of Multimodal Large Language Models: A Survey","ref_index":230,"is_internal_anchor":true},{"citing_arxiv_id":"2604.12115","citing_title":"HTDC: Hesitation-Triggered Differential Calibration for Mitigating Hallucination in Large Vision-Language Models","ref_index":52,"is_internal_anchor":true},{"citing_arxiv_id":"2604.13029","citing_title":"Visual Preference Optimization with Rubric Rewards","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2505.07062","citing_title":"Seed1.5-VL Technical Report","ref_index":180,"is_internal_anchor":true},{"citing_arxiv_id":"2604.18512","citing_title":"S2H-DPO: Hardness-Aware Preference Optimization for Vision-Language Models","ref_index":133,"is_internal_anchor":true},{"citing_arxiv_id":"2604.21911","citing_title":"When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs","ref_index":52,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F","json":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F.json","graph_json":"https://pith.science/api/pith-number/JHTW7AMALJK4JSDR3OR2WJXT3F/graph.json","events_json":"https://pith.science/api/pith-number/JHTW7AMALJK4JSDR3OR2WJXT3F/events.json","paper":"https://pith.science/paper/JHTW7AMA"},"agent_actions":{"view_html":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F","download_json":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F.json","view_paper":"https://pith.science/paper/JHTW7AMA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.11411&json=true","fetch_graph":"https://pith.science/api/pith-number/JHTW7AMALJK4JSDR3OR2WJXT3F/graph.json","fetch_events":"https://pith.science/api/pith-number/JHTW7AMALJK4JSDR3OR2WJXT3F/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F/action/storage_attestation","attest_author":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F/action/author_attestation","sign_citation":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F/action/citation_signature","submit_replication":"https://pith.science/pith/JHTW7AMALJK4JSDR3OR2WJXT3F/action/replication_record"}},"created_at":"2026-05-17T23:38:14.301535+00:00","updated_at":"2026-05-17T23:38:14.301535+00:00"}