{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MQ5DIFWGYUETKECLQS4N57AL46","short_pith_number":"pith:MQ5DIFWG","schema_version":"1.0","canonical_sha256":"643a3416c6c50935104b84b8defc0be79856f1702566e75daf40e90fa95458ae","source":{"kind":"arxiv","id":"2506.07631","version":1},"attestation_state":"computed","paper":{"title":"Unblocking Fine-Grained Evaluation of Detailed Captions: An Explaining AutoRater and Critic-and-Revise Pipeline","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Andreea Marzoca, Brian Gordon, Daniel Cohen-Or, Idan Szpektor, Xiao Wang, Yasumasa Onoe, Yonatan Bitton","submitted_at":"2025-06-09T10:57:26Z","abstract_excerpt":"Large Vision-Language Models (VLMs) now generate highly detailed, paragraphlength image captions, yet evaluating their factual accuracy remains challenging. Current methods often miss fine-grained errors, being designed for shorter texts or lacking datasets with verified inaccuracies. We introduce DOCCI-Critique, a benchmark with 1,400 VLM-generated paragraph captions (100 images, 14 VLMs) featuring over 10,216 sentence-level human annotations of factual correctness and explanatory rationales for errors, all within paragraph context. Building on this, we develop VNLI-Critique, a model for auto"},"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":"2506.07631","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-06-09T10:57:26Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"125200e7a699e3712604ec90574186fffe9468c45b463e1c8b598446e6a788a3","abstract_canon_sha256":"ec71eabdfd454fd31f1b6a3062ad574b266672dcf0ba8f8abde4e75817fac411"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:30.599149Z","signature_b64":"HaKNYfEqVOUD1JHApUsTIzN2BuvS5bnjeCk6vpxuNPsBOXN25ekQrfh+7sr3FH75GvS1y3rMl1O+Sn/kNtS5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"643a3416c6c50935104b84b8defc0be79856f1702566e75daf40e90fa95458ae","last_reissued_at":"2026-07-05T11:18:30.598475Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:30.598475Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unblocking Fine-Grained Evaluation of Detailed Captions: An Explaining AutoRater and Critic-and-Revise Pipeline","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Andreea Marzoca, Brian Gordon, Daniel Cohen-Or, Idan Szpektor, Xiao Wang, Yasumasa Onoe, Yonatan Bitton","submitted_at":"2025-06-09T10:57:26Z","abstract_excerpt":"Large Vision-Language Models (VLMs) now generate highly detailed, paragraphlength image captions, yet evaluating their factual accuracy remains challenging. Current methods often miss fine-grained errors, being designed for shorter texts or lacking datasets with verified inaccuracies. We introduce DOCCI-Critique, a benchmark with 1,400 VLM-generated paragraph captions (100 images, 14 VLMs) featuring over 10,216 sentence-level human annotations of factual correctness and explanatory rationales for errors, all within paragraph context. Building on this, we develop VNLI-Critique, a model for auto"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.07631","kind":"arxiv","version":1},"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/2506.07631/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":"2506.07631","created_at":"2026-07-05T11:18:30.598560+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.07631v1","created_at":"2026-07-05T11:18:30.598560+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.07631","created_at":"2026-07-05T11:18:30.598560+00:00"},{"alias_kind":"pith_short_12","alias_value":"MQ5DIFWGYUET","created_at":"2026-07-05T11:18:30.598560+00:00"},{"alias_kind":"pith_short_16","alias_value":"MQ5DIFWGYUETKECL","created_at":"2026-07-05T11:18:30.598560+00:00"},{"alias_kind":"pith_short_8","alias_value":"MQ5DIFWG","created_at":"2026-07-05T11:18:30.598560+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.21718","citing_title":"Building a Precise Video Language with Human-AI Oversight","ref_index":21,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46","json":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46.json","graph_json":"https://pith.science/api/pith-number/MQ5DIFWGYUETKECLQS4N57AL46/graph.json","events_json":"https://pith.science/api/pith-number/MQ5DIFWGYUETKECLQS4N57AL46/events.json","paper":"https://pith.science/paper/MQ5DIFWG"},"agent_actions":{"view_html":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46","download_json":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46.json","view_paper":"https://pith.science/paper/MQ5DIFWG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.07631&json=true","fetch_graph":"https://pith.science/api/pith-number/MQ5DIFWGYUETKECLQS4N57AL46/graph.json","fetch_events":"https://pith.science/api/pith-number/MQ5DIFWGYUETKECLQS4N57AL46/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46/action/storage_attestation","attest_author":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46/action/author_attestation","sign_citation":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46/action/citation_signature","submit_replication":"https://pith.science/pith/MQ5DIFWGYUETKECLQS4N57AL46/action/replication_record"}},"created_at":"2026-07-05T11:18:30.598560+00:00","updated_at":"2026-07-05T11:18:30.598560+00:00"}