{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:TUGJQA3ILEARTXS3FF5EGONQFF","short_pith_number":"pith:TUGJQA3I","schema_version":"1.0","canonical_sha256":"9d0c980368590119de5b297a4339b02961fdea66a2471422f42f3a3808994aa0","source":{"kind":"arxiv","id":"2204.03162","version":2},"attestation_state":"computed","paper":{"title":"Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Adina Williams, Amanpreet Singh, Candace Ross, Douwe Kiela, Max Bartolo, Ryan Jiang, Tristan Thrush","submitted_at":"2022-04-07T02:17:05Z","abstract_excerpt":"We present a novel task and dataset for evaluating the ability of vision and language models to conduct visio-linguistic compositional reasoning, which we call Winoground. Given two images and two captions, the goal is to match them correctly - but crucially, both captions contain a completely identical set of words, only in a different order. The dataset was carefully hand-curated by expert annotators and is labeled with a rich set of fine-grained tags to assist in analyzing model performance. We probe a diverse range of state-of-the-art vision and language models and find that, surprisingly,"},"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":"2204.03162","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-04-07T02:17:05Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"42db550c621101dec0af33adaedd8745bf8b0133d58240dc122824b6a0941b7c","abstract_canon_sha256":"8802403c62e87647ee3c48887296a29219a9efab635f16ddd5f2abcacb270640"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:17:04.464895Z","signature_b64":"7DKmbEPJCupmzKXKmSoGAfFx2N3qfeEr5p5uJy+ExwDPeYX1a0szlR2dzLXeMc+ePvJaP33hVqcaZ6skrTwNBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9d0c980368590119de5b297a4339b02961fdea66a2471422f42f3a3808994aa0","last_reissued_at":"2026-07-05T04:17:04.464374Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:17:04.464374Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Adina Williams, Amanpreet Singh, Candace Ross, Douwe Kiela, Max Bartolo, Ryan Jiang, Tristan Thrush","submitted_at":"2022-04-07T02:17:05Z","abstract_excerpt":"We present a novel task and dataset for evaluating the ability of vision and language models to conduct visio-linguistic compositional reasoning, which we call Winoground. Given two images and two captions, the goal is to match them correctly - but crucially, both captions contain a completely identical set of words, only in a different order. The dataset was carefully hand-curated by expert annotators and is labeled with a rich set of fine-grained tags to assist in analyzing model performance. We probe a diverse range of state-of-the-art vision and language models and find that, surprisingly,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2204.03162","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/2204.03162/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":"2204.03162","created_at":"2026-07-05T04:17:04.464429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2204.03162v2","created_at":"2026-07-05T04:17:04.464429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2204.03162","created_at":"2026-07-05T04:17:04.464429+00:00"},{"alias_kind":"pith_short_12","alias_value":"TUGJQA3ILEAR","created_at":"2026-07-05T04:17:04.464429+00:00"},{"alias_kind":"pith_short_16","alias_value":"TUGJQA3ILEARTXS3","created_at":"2026-07-05T04:17:04.464429+00:00"},{"alias_kind":"pith_short_8","alias_value":"TUGJQA3I","created_at":"2026-07-05T04:17:04.464429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2505.23678","citing_title":"Grounded Reinforcement Learning for Visual Reasoning","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05810","citing_title":"CXR-ContraBench: Benchmarking Negated-Option Attraction in Medical VLMs","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF","json":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF.json","graph_json":"https://pith.science/api/pith-number/TUGJQA3ILEARTXS3FF5EGONQFF/graph.json","events_json":"https://pith.science/api/pith-number/TUGJQA3ILEARTXS3FF5EGONQFF/events.json","paper":"https://pith.science/paper/TUGJQA3I"},"agent_actions":{"view_html":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF","download_json":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF.json","view_paper":"https://pith.science/paper/TUGJQA3I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2204.03162&json=true","fetch_graph":"https://pith.science/api/pith-number/TUGJQA3ILEARTXS3FF5EGONQFF/graph.json","fetch_events":"https://pith.science/api/pith-number/TUGJQA3ILEARTXS3FF5EGONQFF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF/action/storage_attestation","attest_author":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF/action/author_attestation","sign_citation":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF/action/citation_signature","submit_replication":"https://pith.science/pith/TUGJQA3ILEARTXS3FF5EGONQFF/action/replication_record"}},"created_at":"2026-07-05T04:17:04.464429+00:00","updated_at":"2026-07-05T04:17:04.464429+00:00"}