{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YFX7MQYY2CCUNDRCRO54NMJPEN","short_pith_number":"pith:YFX7MQYY","schema_version":"1.0","canonical_sha256":"c16ff64318d085468e228bbbc6b12f23632347c3aa83bcb887b770cfd8112154","source":{"kind":"arxiv","id":"2505.04673","version":1},"attestation_state":"computed","paper":{"title":"REVEAL: Multi-turn Evaluation of Image-Input Harms for Vision LLM","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Madhur Jindal, Saurabh Deshpande","submitted_at":"2025-05-07T10:09:55Z","abstract_excerpt":"Vision Large Language Models (VLLMs) represent a significant advancement in artificial intelligence by integrating image-processing capabilities with textual understanding, thereby enhancing user interactions and expanding application domains. However, their increased complexity introduces novel safety and ethical challenges, particularly in multi-modal and multi-turn conversations. Traditional safety evaluation frameworks, designed for text-based, single-turn interactions, are inadequate for addressing these complexities. To bridge this gap, we introduce the REVEAL (Responsible Evaluation of "},"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":"2505.04673","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-07T10:09:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"693a4f65533ac5caa41c7cd00080a169d75b500fdf44d0f9cac12630afabe757","abstract_canon_sha256":"fee9fd1d95835ba2b52f828c0058cd64f873069c67203832d46c2e2ab3def709"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:11.572027Z","signature_b64":"ptn1RD7ZC0wSGPGEQcxtW5ADM8EIGM8x6yh/tyfTtoGU8U0vM1IwPSXJ0u4A9RMae4fQpysdFZTJQ/cTjJmsDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c16ff64318d085468e228bbbc6b12f23632347c3aa83bcb887b770cfd8112154","last_reissued_at":"2026-07-05T11:00:11.571576Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:11.571576Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REVEAL: Multi-turn Evaluation of Image-Input Harms for Vision LLM","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Madhur Jindal, Saurabh Deshpande","submitted_at":"2025-05-07T10:09:55Z","abstract_excerpt":"Vision Large Language Models (VLLMs) represent a significant advancement in artificial intelligence by integrating image-processing capabilities with textual understanding, thereby enhancing user interactions and expanding application domains. However, their increased complexity introduces novel safety and ethical challenges, particularly in multi-modal and multi-turn conversations. Traditional safety evaluation frameworks, designed for text-based, single-turn interactions, are inadequate for addressing these complexities. To bridge this gap, we introduce the REVEAL (Responsible Evaluation of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.04673","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/2505.04673/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":"2505.04673","created_at":"2026-07-05T11:00:11.571637+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.04673v1","created_at":"2026-07-05T11:00:11.571637+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.04673","created_at":"2026-07-05T11:00:11.571637+00:00"},{"alias_kind":"pith_short_12","alias_value":"YFX7MQYY2CCU","created_at":"2026-07-05T11:00:11.571637+00:00"},{"alias_kind":"pith_short_16","alias_value":"YFX7MQYY2CCUNDRC","created_at":"2026-07-05T11:00:11.571637+00:00"},{"alias_kind":"pith_short_8","alias_value":"YFX7MQYY","created_at":"2026-07-05T11:00:11.571637+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.24082","citing_title":"Jailbreaking Frontier Foundation Models Through Intention Deception","ref_index":4,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN","json":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN.json","graph_json":"https://pith.science/api/pith-number/YFX7MQYY2CCUNDRCRO54NMJPEN/graph.json","events_json":"https://pith.science/api/pith-number/YFX7MQYY2CCUNDRCRO54NMJPEN/events.json","paper":"https://pith.science/paper/YFX7MQYY"},"agent_actions":{"view_html":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN","download_json":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN.json","view_paper":"https://pith.science/paper/YFX7MQYY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.04673&json=true","fetch_graph":"https://pith.science/api/pith-number/YFX7MQYY2CCUNDRCRO54NMJPEN/graph.json","fetch_events":"https://pith.science/api/pith-number/YFX7MQYY2CCUNDRCRO54NMJPEN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN/action/storage_attestation","attest_author":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN/action/author_attestation","sign_citation":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN/action/citation_signature","submit_replication":"https://pith.science/pith/YFX7MQYY2CCUNDRCRO54NMJPEN/action/replication_record"}},"created_at":"2026-07-05T11:00:11.571637+00:00","updated_at":"2026-07-05T11:00:11.571637+00:00"}