{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:A6QQDHT3U2N3HTIZGC37YER76G","short_pith_number":"pith:A6QQDHT3","schema_version":"1.0","canonical_sha256":"07a1019e7ba69bb3cd1930b7fc123ff1a4e3f82e7a1cb6dcbb13faa5b1cd1ada","source":{"kind":"arxiv","id":"2412.15484","version":4},"attestation_state":"computed","paper":{"title":"Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jing Shi, Saehyung Lee, Seunghyun Yoon, Sungroh Yoon, Trung Bui","submitted_at":"2024-12-20T01:37:22Z","abstract_excerpt":"Multimodal large language models (MLLMs) excel at generating highly detailed captions but often produce hallucinations. Our analysis reveals that existing hallucination detection methods struggle with detailed captions. We attribute this to the increasing reliance of MLLMs on their generated text, rather than the input image, as the sequence length grows. To address this issue, we propose a multiagent approach that leverages LLM-MLLM collaboration to correct given captions. Additionally, we introduce an evaluation framework and a benchmark dataset to facilitate the systematic analysis of detai"},"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":"2412.15484","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-20T01:37:22Z","cross_cats_sorted":[],"title_canon_sha256":"7d9a45c95cabeea2338b5abe6152e0e36df294faf6416d888e4593c27773dc9f","abstract_canon_sha256":"830c401c3724c8d9ddf497da8f9277ca611142203e921b77569d13e70cbfd13b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:32:22.029147Z","signature_b64":"WXXqrO0maR68DdgdmJy4sBZ1rpzuB3Fi2wAezKx4hSb6CuSZqeSyVxwzCUryG8BfgxlwZ3rqBB/1N+zpsKhQAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"07a1019e7ba69bb3cd1930b7fc123ff1a4e3f82e7a1cb6dcbb13faa5b1cd1ada","last_reissued_at":"2026-07-05T11:32:22.028655Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:32:22.028655Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jing Shi, Saehyung Lee, Seunghyun Yoon, Sungroh Yoon, Trung Bui","submitted_at":"2024-12-20T01:37:22Z","abstract_excerpt":"Multimodal large language models (MLLMs) excel at generating highly detailed captions but often produce hallucinations. Our analysis reveals that existing hallucination detection methods struggle with detailed captions. We attribute this to the increasing reliance of MLLMs on their generated text, rather than the input image, as the sequence length grows. To address this issue, we propose a multiagent approach that leverages LLM-MLLM collaboration to correct given captions. Additionally, we introduce an evaluation framework and a benchmark dataset to facilitate the systematic analysis of detai"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.15484","kind":"arxiv","version":4},"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/2412.15484/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":"2412.15484","created_at":"2026-07-05T11:32:22.028707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.15484v4","created_at":"2026-07-05T11:32:22.028707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.15484","created_at":"2026-07-05T11:32:22.028707+00:00"},{"alias_kind":"pith_short_12","alias_value":"A6QQDHT3U2N3","created_at":"2026-07-05T11:32:22.028707+00:00"},{"alias_kind":"pith_short_16","alias_value":"A6QQDHT3U2N3HTIZ","created_at":"2026-07-05T11:32:22.028707+00:00"},{"alias_kind":"pith_short_8","alias_value":"A6QQDHT3","created_at":"2026-07-05T11:32:22.028707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09393","citing_title":"CapRL++: Unified Reinforcement Learning with Verifiable Rewards for Dense Image and Video Captioning","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07647","citing_title":"Steer Where It Matters: Token-Level Visual-Sensitivity Steering for LVLMs Hallucination Mitigation","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G","json":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G.json","graph_json":"https://pith.science/api/pith-number/A6QQDHT3U2N3HTIZGC37YER76G/graph.json","events_json":"https://pith.science/api/pith-number/A6QQDHT3U2N3HTIZGC37YER76G/events.json","paper":"https://pith.science/paper/A6QQDHT3"},"agent_actions":{"view_html":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G","download_json":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G.json","view_paper":"https://pith.science/paper/A6QQDHT3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.15484&json=true","fetch_graph":"https://pith.science/api/pith-number/A6QQDHT3U2N3HTIZGC37YER76G/graph.json","fetch_events":"https://pith.science/api/pith-number/A6QQDHT3U2N3HTIZGC37YER76G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G/action/storage_attestation","attest_author":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G/action/author_attestation","sign_citation":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G/action/citation_signature","submit_replication":"https://pith.science/pith/A6QQDHT3U2N3HTIZGC37YER76G/action/replication_record"}},"created_at":"2026-07-05T11:32:22.028707+00:00","updated_at":"2026-07-05T11:32:22.028707+00:00"}