{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SH2GEG5J3TUDUM7ZVGZD2UZMJE","short_pith_number":"pith:SH2GEG5J","schema_version":"1.0","canonical_sha256":"91f4621ba9dce83a33f9a9b23d532c492e4e7a34a85bafcba98e83be0e0c2641","source":{"kind":"arxiv","id":"2411.02708","version":3},"attestation_state":"computed","paper":{"title":"Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Aiwei Liu, Jia Liu, Jingyu Wang, Jungang Li, Mengxi Gao, Peijie Jiang, Xin Zou, Xuming Hu, Yanggan Gu, Yibo Yan, Yunkai Dang","submitted_at":"2024-11-05T01:11:28Z","abstract_excerpt":"Multimodal large language models (MLLMs) have recently achieved state-of-the-art performance on tasks ranging from visual question answering to video understanding. However, existing studies have concentrated mainly on visual-textual misalignment, leaving largely unexplored the MLLMs' ability to preserve an originally correct answer when confronted with misleading information. We reveal a response uncertainty phenomenon: across nine standard datasets, twelve state-of-the-art open-source MLLMs overturn a previously correct answer in 65% of cases after receiving a single deceptive cue. To system"},"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":"2411.02708","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-05T01:11:28Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"e0f2c2e42d3e02b7c3fe90309e10b1bdc49009885d68acbedae5019efdc72865","abstract_canon_sha256":"aa0a42d50394a47b6b43ab8b8fb05b5aa68831882dc7f644cb0c4a748f181016"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:03:44.021414Z","signature_b64":"IZs2RiLYnnYDVfnW58wi3/+SBMTIVNvcUHuplU2+DeHRNgAtvseEm8lWpA4LN/j8bDcEi4y5RMaEqsUL3deiBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91f4621ba9dce83a33f9a9b23d532c492e4e7a34a85bafcba98e83be0e0c2641","last_reissued_at":"2026-07-05T12:03:44.020879Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:03:44.020879Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exploring Response Uncertainty in MLLMs: An Empirical Evaluation under Misleading Scenarios","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Aiwei Liu, Jia Liu, Jingyu Wang, Jungang Li, Mengxi Gao, Peijie Jiang, Xin Zou, Xuming Hu, Yanggan Gu, Yibo Yan, Yunkai Dang","submitted_at":"2024-11-05T01:11:28Z","abstract_excerpt":"Multimodal large language models (MLLMs) have recently achieved state-of-the-art performance on tasks ranging from visual question answering to video understanding. However, existing studies have concentrated mainly on visual-textual misalignment, leaving largely unexplored the MLLMs' ability to preserve an originally correct answer when confronted with misleading information. We reveal a response uncertainty phenomenon: across nine standard datasets, twelve state-of-the-art open-source MLLMs overturn a previously correct answer in 65% of cases after receiving a single deceptive cue. To system"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.02708","kind":"arxiv","version":3},"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/2411.02708/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":"2411.02708","created_at":"2026-07-05T12:03:44.020946+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.02708v3","created_at":"2026-07-05T12:03:44.020946+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.02708","created_at":"2026-07-05T12:03:44.020946+00:00"},{"alias_kind":"pith_short_12","alias_value":"SH2GEG5J3TUD","created_at":"2026-07-05T12:03:44.020946+00:00"},{"alias_kind":"pith_short_16","alias_value":"SH2GEG5J3TUDUM7Z","created_at":"2026-07-05T12:03:44.020946+00:00"},{"alias_kind":"pith_short_8","alias_value":"SH2GEG5J","created_at":"2026-07-05T12:03:44.020946+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.02871","citing_title":"Position: Multimodal Large Language Models Can Significantly Advance Scientific Reasoning","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2511.14159","citing_title":"MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE","json":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE.json","graph_json":"https://pith.science/api/pith-number/SH2GEG5J3TUDUM7ZVGZD2UZMJE/graph.json","events_json":"https://pith.science/api/pith-number/SH2GEG5J3TUDUM7ZVGZD2UZMJE/events.json","paper":"https://pith.science/paper/SH2GEG5J"},"agent_actions":{"view_html":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE","download_json":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE.json","view_paper":"https://pith.science/paper/SH2GEG5J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.02708&json=true","fetch_graph":"https://pith.science/api/pith-number/SH2GEG5J3TUDUM7ZVGZD2UZMJE/graph.json","fetch_events":"https://pith.science/api/pith-number/SH2GEG5J3TUDUM7ZVGZD2UZMJE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE/action/storage_attestation","attest_author":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE/action/author_attestation","sign_citation":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE/action/citation_signature","submit_replication":"https://pith.science/pith/SH2GEG5J3TUDUM7ZVGZD2UZMJE/action/replication_record"}},"created_at":"2026-07-05T12:03:44.020946+00:00","updated_at":"2026-07-05T12:03:44.020946+00:00"}