{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:66DQDQ2RCYBFVFF2IXRHAITLOX","short_pith_number":"pith:66DQDQ2R","schema_version":"1.0","canonical_sha256":"f78701c35116025a94ba45e270226b75cd503b69b0b146b7c5ac4e9e2eef84f9","source":{"kind":"arxiv","id":"2501.09425","version":2},"attestation_state":"computed","paper":{"title":"Vision-Language Models Do Not Understand Negation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Guohao Li, Kumail Alhamoud, Marzyeh Ghassemi, Philip Torr, Shaden Alshammari, Yonglong Tian, Yoon Kim","submitted_at":"2025-01-16T09:55:42Z","abstract_excerpt":"Many practical vision-language applications require models that understand negation, e.g., when using natural language to retrieve images which contain certain objects but not others. Despite advancements in vision-language models (VLMs) through large-scale training, their ability to comprehend negation remains underexplored. This study addresses the question: how well do current VLMs understand negation? We introduce NegBench, a new benchmark designed to evaluate negation understanding across 18 task variations and $79$k examples spanning image, video, and medical datasets. The benchmark cons"},"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":"2501.09425","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-01-16T09:55:42Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"bba76a857d7b2c5f0dc0a80fe0df7a0a96c9d248d4b8c8c6312a792de1be8e7d","abstract_canon_sha256":"f1359a2da90e2dcea711a89ab88c6a930bf92986b03ed0da92c716808d7fabd1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:02:16.710863Z","signature_b64":"BO4mT+vzZjPtwl8E/ZuXdxkn+Mb/6zqW3NP8qF4QSgl6ndJ3CVjjarleJ54Gx8Xa/phnbLawaLFLlewwbmYZAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f78701c35116025a94ba45e270226b75cd503b69b0b146b7c5ac4e9e2eef84f9","last_reissued_at":"2026-07-05T11:02:16.710218Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:02:16.710218Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Vision-Language Models Do Not Understand Negation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Guohao Li, Kumail Alhamoud, Marzyeh Ghassemi, Philip Torr, Shaden Alshammari, Yonglong Tian, Yoon Kim","submitted_at":"2025-01-16T09:55:42Z","abstract_excerpt":"Many practical vision-language applications require models that understand negation, e.g., when using natural language to retrieve images which contain certain objects but not others. Despite advancements in vision-language models (VLMs) through large-scale training, their ability to comprehend negation remains underexplored. This study addresses the question: how well do current VLMs understand negation? We introduce NegBench, a new benchmark designed to evaluate negation understanding across 18 task variations and $79$k examples spanning image, video, and medical datasets. The benchmark cons"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.09425","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/2501.09425/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":"2501.09425","created_at":"2026-07-05T11:02:16.710291+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.09425v2","created_at":"2026-07-05T11:02:16.710291+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.09425","created_at":"2026-07-05T11:02:16.710291+00:00"},{"alias_kind":"pith_short_12","alias_value":"66DQDQ2RCYBF","created_at":"2026-07-05T11:02:16.710291+00:00"},{"alias_kind":"pith_short_16","alias_value":"66DQDQ2RCYBFVFF2","created_at":"2026-07-05T11:02:16.710291+00:00"},{"alias_kind":"pith_short_8","alias_value":"66DQDQ2R","created_at":"2026-07-05T11:02:16.710291+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.22773","citing_title":"Trace Mutation in Human-LLM Dialogue: The Transcript as Forensic and Mitigation Surface","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05810","citing_title":"CXR-ContraBench: Benchmarking Negated-Option Attraction in Medical VLMs","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18942","citing_title":"Disparities In Negation Understanding Across Languages In Vision-Language Models","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06815","citing_title":"Uneven Evolution of Cognition Across Generations of Generative AI Models","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX","json":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX.json","graph_json":"https://pith.science/api/pith-number/66DQDQ2RCYBFVFF2IXRHAITLOX/graph.json","events_json":"https://pith.science/api/pith-number/66DQDQ2RCYBFVFF2IXRHAITLOX/events.json","paper":"https://pith.science/paper/66DQDQ2R"},"agent_actions":{"view_html":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX","download_json":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX.json","view_paper":"https://pith.science/paper/66DQDQ2R","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.09425&json=true","fetch_graph":"https://pith.science/api/pith-number/66DQDQ2RCYBFVFF2IXRHAITLOX/graph.json","fetch_events":"https://pith.science/api/pith-number/66DQDQ2RCYBFVFF2IXRHAITLOX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX/action/storage_attestation","attest_author":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX/action/author_attestation","sign_citation":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX/action/citation_signature","submit_replication":"https://pith.science/pith/66DQDQ2RCYBFVFF2IXRHAITLOX/action/replication_record"}},"created_at":"2026-07-05T11:02:16.710291+00:00","updated_at":"2026-07-05T11:02:16.710291+00:00"}