{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5OOGL74WVCCLTLBCBN3K4SIDQY","short_pith_number":"pith:5OOGL74W","schema_version":"1.0","canonical_sha256":"eb9c65ff96a884b9ac220b76ae49038626a090952482551c697a6c296db1e270","source":{"kind":"arxiv","id":"2412.02172","version":2},"attestation_state":"computed","paper":{"title":"VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Bingxuan Li, Da Yin, Kai-Wei Chang, Nanyun Peng, Pan Lu, Xueqing Wu, Yuheng Ding","submitted_at":"2024-12-03T05:04:49Z","abstract_excerpt":"The ability of large vision-language models (LVLMs) to critique and correct their reasoning is an essential building block towards their self-improvement. However, a systematic analysis of such capabilities in LVLMs is still lacking. We propose VISCO, the first benchmark to extensively analyze the fine-grained critique and correction capabilities of LVLMs. Compared to existing work that uses a single scalar value to critique the entire reasoning [4], VISCO features dense and fine-grained critique, requiring LVLMs to evaluate the correctness of each step in the chain-of-thought and provide natu"},"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":"2412.02172","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-03T05:04:49Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"83109f974393ed215036f2091aa1c6d41266a92b437cc068eaeab0e6197bf67f","abstract_canon_sha256":"40b532dac5948d451386a9e7d9803b367fbf227ef69b024c07957c3e6b9a2ccf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:33:40.628922Z","signature_b64":"jrjNLBnNP+rfCYnnCq9foVJKHIqoHEWNI4CY7uIQWEXjoWlAB7HWJ9mCnqIDGn3Zhap0ydtDZzdK4KBCBz9dCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eb9c65ff96a884b9ac220b76ae49038626a090952482551c697a6c296db1e270","last_reissued_at":"2026-07-05T10:33:40.628407Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:33:40.628407Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VISCO: Benchmarking Fine-Grained Critique and Correction Towards Self-Improvement in Visual Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Bingxuan Li, Da Yin, Kai-Wei Chang, Nanyun Peng, Pan Lu, Xueqing Wu, Yuheng Ding","submitted_at":"2024-12-03T05:04:49Z","abstract_excerpt":"The ability of large vision-language models (LVLMs) to critique and correct their reasoning is an essential building block towards their self-improvement. However, a systematic analysis of such capabilities in LVLMs is still lacking. We propose VISCO, the first benchmark to extensively analyze the fine-grained critique and correction capabilities of LVLMs. Compared to existing work that uses a single scalar value to critique the entire reasoning [4], VISCO features dense and fine-grained critique, requiring LVLMs to evaluate the correctness of each step in the chain-of-thought and provide natu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.02172","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/2412.02172/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":"2412.02172","created_at":"2026-07-05T10:33:40.628465+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.02172v2","created_at":"2026-07-05T10:33:40.628465+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.02172","created_at":"2026-07-05T10:33:40.628465+00:00"},{"alias_kind":"pith_short_12","alias_value":"5OOGL74WVCCL","created_at":"2026-07-05T10:33:40.628465+00:00"},{"alias_kind":"pith_short_16","alias_value":"5OOGL74WVCCLTLBC","created_at":"2026-07-05T10:33:40.628465+00:00"},{"alias_kind":"pith_short_8","alias_value":"5OOGL74W","created_at":"2026-07-05T10:33:40.628465+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.07297","citing_title":"MagiC: Evaluating Multimodal Cognition Toward Grounded Visual Reasoning","ref_index":31,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY","json":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY.json","graph_json":"https://pith.science/api/pith-number/5OOGL74WVCCLTLBCBN3K4SIDQY/graph.json","events_json":"https://pith.science/api/pith-number/5OOGL74WVCCLTLBCBN3K4SIDQY/events.json","paper":"https://pith.science/paper/5OOGL74W"},"agent_actions":{"view_html":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY","download_json":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY.json","view_paper":"https://pith.science/paper/5OOGL74W","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.02172&json=true","fetch_graph":"https://pith.science/api/pith-number/5OOGL74WVCCLTLBCBN3K4SIDQY/graph.json","fetch_events":"https://pith.science/api/pith-number/5OOGL74WVCCLTLBCBN3K4SIDQY/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY/action/storage_attestation","attest_author":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY/action/author_attestation","sign_citation":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY/action/citation_signature","submit_replication":"https://pith.science/pith/5OOGL74WVCCLTLBCBN3K4SIDQY/action/replication_record"}},"created_at":"2026-07-05T10:33:40.628465+00:00","updated_at":"2026-07-05T10:33:40.628465+00:00"}