{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:X2X5FRFBISAICU3AY3UMAIX3BW","short_pith_number":"pith:X2X5FRFB","schema_version":"1.0","canonical_sha256":"beafd2c4a14480815360c6e8c022fb0da866aa5a55d77afe4cac652212bcf671","source":{"kind":"arxiv","id":"2411.08410","version":2},"attestation_state":"computed","paper":{"title":"The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defense","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CR","authors_text":"Fangkai Jiao, Liqiang Nie, Mohan Kankanhalli, Yangyang Guo","submitted_at":"2024-11-13T07:57:19Z","abstract_excerpt":"The vulnerability of Vision Large Language Models (VLLMs) to jailbreak attacks appears as no surprise. However, recent defense mechanisms against these attacks have reached near-saturation performance on benchmark evaluations, often with minimal effort. This \\emph{dual high performance} in both attack and defense raises a fundamental and perplexing paradox. To gain a deep understanding of this issue and thus further help strengthen the trustworthiness of VLLMs, this paper makes three key contributions: i) One tentative explanation for VLLMs being prone to jailbreak attacks--\\textbf{inclusion o"},"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":"2411.08410","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-11-13T07:57:19Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"ac0dec5b050477464a239ef9df5ab6303e3b6be60463a2b14e183723f2a07a95","abstract_canon_sha256":"f16eb665156eaf0c185297215263c9734a0c0aa3be59d301577f2e650a5c2263"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:25:13.744511Z","signature_b64":"/kBp1rwagfMv6Itn71/dmsNslDaAkzGWzdjfj7wpqR0OigwWXW//Sk2rhuzd9z/yfM5Ozd+8/zSNVgNjl867CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"beafd2c4a14480815360c6e8c022fb0da866aa5a55d77afe4cac652212bcf671","last_reissued_at":"2026-07-05T10:25:13.743668Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:25:13.743668Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The VLLM Safety Paradox: Dual Ease in Jailbreak Attack and Defense","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CR","authors_text":"Fangkai Jiao, Liqiang Nie, Mohan Kankanhalli, Yangyang Guo","submitted_at":"2024-11-13T07:57:19Z","abstract_excerpt":"The vulnerability of Vision Large Language Models (VLLMs) to jailbreak attacks appears as no surprise. However, recent defense mechanisms against these attacks have reached near-saturation performance on benchmark evaluations, often with minimal effort. This \\emph{dual high performance} in both attack and defense raises a fundamental and perplexing paradox. To gain a deep understanding of this issue and thus further help strengthen the trustworthiness of VLLMs, this paper makes three key contributions: i) One tentative explanation for VLLMs being prone to jailbreak attacks--\\textbf{inclusion o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.08410","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/2411.08410/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":"2411.08410","created_at":"2026-07-05T10:25:13.743762+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.08410v2","created_at":"2026-07-05T10:25:13.743762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.08410","created_at":"2026-07-05T10:25:13.743762+00:00"},{"alias_kind":"pith_short_12","alias_value":"X2X5FRFBISAI","created_at":"2026-07-05T10:25:13.743762+00:00"},{"alias_kind":"pith_short_16","alias_value":"X2X5FRFBISAICU3A","created_at":"2026-07-05T10:25:13.743762+00:00"},{"alias_kind":"pith_short_8","alias_value":"X2X5FRFB","created_at":"2026-07-05T10:25:13.743762+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2409.18169","citing_title":"Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW","json":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW.json","graph_json":"https://pith.science/api/pith-number/X2X5FRFBISAICU3AY3UMAIX3BW/graph.json","events_json":"https://pith.science/api/pith-number/X2X5FRFBISAICU3AY3UMAIX3BW/events.json","paper":"https://pith.science/paper/X2X5FRFB"},"agent_actions":{"view_html":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW","download_json":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW.json","view_paper":"https://pith.science/paper/X2X5FRFB","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.08410&json=true","fetch_graph":"https://pith.science/api/pith-number/X2X5FRFBISAICU3AY3UMAIX3BW/graph.json","fetch_events":"https://pith.science/api/pith-number/X2X5FRFBISAICU3AY3UMAIX3BW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW/action/storage_attestation","attest_author":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW/action/author_attestation","sign_citation":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW/action/citation_signature","submit_replication":"https://pith.science/pith/X2X5FRFBISAICU3AY3UMAIX3BW/action/replication_record"}},"created_at":"2026-07-05T10:25:13.743762+00:00","updated_at":"2026-07-05T10:25:13.743762+00:00"}