{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4WW2N7EW4NJVBRW3OPVX33FBJM","short_pith_number":"pith:4WW2N7EW","schema_version":"1.0","canonical_sha256":"e5ada6fc96e35350c6db73eb7deca14b10391335f7a14bd98b3015a5e236b5d8","source":{"kind":"arxiv","id":"2403.18346","version":4},"attestation_state":"computed","paper":{"title":"Quantifying and Mitigating Unimodal Biases in Multimodal Large Language Models: A Causal Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Chaochao Lu, Meiqi Chen, Yan Zhang, Yixin Cao","submitted_at":"2024-03-27T08:38:49Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have facilitated the development of Multimodal LLMs (MLLMs). Despite their impressive capabilities, MLLMs often suffer from over-reliance on unimodal biases (e.g., language bias and vision bias), leading to incorrect answers or hallucinations in complex multimodal tasks. To investigate this issue, we propose a causal framework to interpret the biases in Visual Question Answering (VQA) problems. Within this framework, we conduct an in-depth causal analysis to assess the causal effect of these biases on MLLM predictions. Based on the analysis, "},"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":"2403.18346","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-03-27T08:38:49Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"214b982f474b7c7b2452d8f7e624241451dec7971b484856498184d12ee5692f","abstract_canon_sha256":"df91b132c6f92127540b4abdcc7d213aedda87a3f1be98ae323c49fc234182e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:34:56.248488Z","signature_b64":"yX4ZnMfWXfY8dvYonRad8RD5gE4scecx9DlAiPN9cZhL5BYpInAvYZPZ3Z03grA1wljHM+qVoTDB24hifxztCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e5ada6fc96e35350c6db73eb7deca14b10391335f7a14bd98b3015a5e236b5d8","last_reissued_at":"2026-07-05T09:34:56.247913Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:34:56.247913Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Quantifying and Mitigating Unimodal Biases in Multimodal Large Language Models: A Causal Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Chaochao Lu, Meiqi Chen, Yan Zhang, Yixin Cao","submitted_at":"2024-03-27T08:38:49Z","abstract_excerpt":"Recent advancements in Large Language Models (LLMs) have facilitated the development of Multimodal LLMs (MLLMs). Despite their impressive capabilities, MLLMs often suffer from over-reliance on unimodal biases (e.g., language bias and vision bias), leading to incorrect answers or hallucinations in complex multimodal tasks. To investigate this issue, we propose a causal framework to interpret the biases in Visual Question Answering (VQA) problems. Within this framework, we conduct an in-depth causal analysis to assess the causal effect of these biases on MLLM predictions. Based on the analysis, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.18346","kind":"arxiv","version":4},"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/2403.18346/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":"2403.18346","created_at":"2026-07-05T09:34:56.247969+00:00"},{"alias_kind":"arxiv_version","alias_value":"2403.18346v4","created_at":"2026-07-05T09:34:56.247969+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.18346","created_at":"2026-07-05T09:34:56.247969+00:00"},{"alias_kind":"pith_short_12","alias_value":"4WW2N7EW4NJV","created_at":"2026-07-05T09:34:56.247969+00:00"},{"alias_kind":"pith_short_16","alias_value":"4WW2N7EW4NJVBRW3","created_at":"2026-07-05T09:34:56.247969+00:00"},{"alias_kind":"pith_short_8","alias_value":"4WW2N7EW","created_at":"2026-07-05T09:34:56.247969+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.18043","citing_title":"GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs","ref_index":6,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM","json":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM.json","graph_json":"https://pith.science/api/pith-number/4WW2N7EW4NJVBRW3OPVX33FBJM/graph.json","events_json":"https://pith.science/api/pith-number/4WW2N7EW4NJVBRW3OPVX33FBJM/events.json","paper":"https://pith.science/paper/4WW2N7EW"},"agent_actions":{"view_html":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM","download_json":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM.json","view_paper":"https://pith.science/paper/4WW2N7EW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2403.18346&json=true","fetch_graph":"https://pith.science/api/pith-number/4WW2N7EW4NJVBRW3OPVX33FBJM/graph.json","fetch_events":"https://pith.science/api/pith-number/4WW2N7EW4NJVBRW3OPVX33FBJM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM/action/storage_attestation","attest_author":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM/action/author_attestation","sign_citation":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM/action/citation_signature","submit_replication":"https://pith.science/pith/4WW2N7EW4NJVBRW3OPVX33FBJM/action/replication_record"}},"created_at":"2026-07-05T09:34:56.247969+00:00","updated_at":"2026-07-05T09:34:56.247969+00:00"}