{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4F4VIJEDYTFEIHB5RO4F7VCCB5","short_pith_number":"pith:4F4VIJED","schema_version":"1.0","canonical_sha256":"e179542483c4ca441c3d8bb85fd4420f563bb51382385f8e9dc7826b3f392127","source":{"kind":"arxiv","id":"2405.05256","version":2},"attestation_state":"computed","paper":{"title":"THRONE: An Object-based Hallucination Benchmark for the Free-form Generations of Large Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Ashwin Swaminathan, C. J. Taylor, Hao Yang, Prannay Kaul, Stefano Soatto, Yonatan Dukler, Zhizhong Li","submitted_at":"2024-05-08T17:59:11Z","abstract_excerpt":"Mitigating hallucinations in large vision-language models (LVLMs) remains an open problem. Recent benchmarks do not address hallucinations in open-ended free-form responses, which we term \"Type I hallucinations\". Instead, they focus on hallucinations responding to very specific question formats -- typically a multiple-choice response regarding a particular object or attribute -- which we term \"Type II hallucinations\". Additionally, such benchmarks often require external API calls to models which are subject to change. In practice, we observe that a reduction in Type II hallucinations does not "},"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":"2405.05256","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-05-08T17:59:11Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"78f5efe086d274767e43d808a53197bc1f3ff8e1d554c81e12c93c87faa4a3a4","abstract_canon_sha256":"5359401e3a43367137b82b9d94ce3b60048ee30676d5bf40ad6094948994e000"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:43:35.674271Z","signature_b64":"FABcweJqnAdzYo2URTAXv8kG6Lx/G+1A3ys/u+8ZSaOWuQ/G2PbkzvfeM3hgG3U11jaDHm75PgWYIPTH1Kb5DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e179542483c4ca441c3d8bb85fd4420f563bb51382385f8e9dc7826b3f392127","last_reissued_at":"2026-07-05T10:43:35.673819Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:43:35.673819Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"THRONE: An Object-based Hallucination Benchmark for the Free-form Generations of Large Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Ashwin Swaminathan, C. J. Taylor, Hao Yang, Prannay Kaul, Stefano Soatto, Yonatan Dukler, Zhizhong Li","submitted_at":"2024-05-08T17:59:11Z","abstract_excerpt":"Mitigating hallucinations in large vision-language models (LVLMs) remains an open problem. Recent benchmarks do not address hallucinations in open-ended free-form responses, which we term \"Type I hallucinations\". Instead, they focus on hallucinations responding to very specific question formats -- typically a multiple-choice response regarding a particular object or attribute -- which we term \"Type II hallucinations\". Additionally, such benchmarks often require external API calls to models which are subject to change. In practice, we observe that a reduction in Type II hallucinations does not "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.05256","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/2405.05256/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":"2405.05256","created_at":"2026-07-05T10:43:35.673877+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.05256v2","created_at":"2026-07-05T10:43:35.673877+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.05256","created_at":"2026-07-05T10:43:35.673877+00:00"},{"alias_kind":"pith_short_12","alias_value":"4F4VIJEDYTFE","created_at":"2026-07-05T10:43:35.673877+00:00"},{"alias_kind":"pith_short_16","alias_value":"4F4VIJEDYTFEIHB5","created_at":"2026-07-05T10:43:35.673877+00:00"},{"alias_kind":"pith_short_8","alias_value":"4F4VIJED","created_at":"2026-07-05T10:43:35.673877+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.06165","citing_title":"HaloProbe: Bayesian Detection and Mitigation of Object Hallucinations in Vision-Language Models","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5","json":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5.json","graph_json":"https://pith.science/api/pith-number/4F4VIJEDYTFEIHB5RO4F7VCCB5/graph.json","events_json":"https://pith.science/api/pith-number/4F4VIJEDYTFEIHB5RO4F7VCCB5/events.json","paper":"https://pith.science/paper/4F4VIJED"},"agent_actions":{"view_html":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5","download_json":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5.json","view_paper":"https://pith.science/paper/4F4VIJED","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.05256&json=true","fetch_graph":"https://pith.science/api/pith-number/4F4VIJEDYTFEIHB5RO4F7VCCB5/graph.json","fetch_events":"https://pith.science/api/pith-number/4F4VIJEDYTFEIHB5RO4F7VCCB5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5/action/storage_attestation","attest_author":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5/action/author_attestation","sign_citation":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5/action/citation_signature","submit_replication":"https://pith.science/pith/4F4VIJEDYTFEIHB5RO4F7VCCB5/action/replication_record"}},"created_at":"2026-07-05T10:43:35.673877+00:00","updated_at":"2026-07-05T10:43:35.673877+00:00"}