{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:MIKHFJIMYJFZ6XVWURUGY2RUTJ","short_pith_number":"pith:MIKHFJIM","schema_version":"1.0","canonical_sha256":"621472a50cc24b9f5eb6a4686c6a349a689c0f9eac49be6ae0cf8c2d3f48ced8","source":{"kind":"arxiv","id":"2412.05934","version":3},"attestation_state":"computed","paper":{"title":"Heuristic-Induced Multimodal Risk Distribution Jailbreak Attack for Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Chu Zhixuan, Duan Ranjie, Huang Yihao, Jia Xiaojun, Jia Xiaoshuang, Li Xinfeng, Ma Teng, Ren Wenqi","submitted_at":"2024-12-08T13:20:45Z","abstract_excerpt":"With the rapid advancement of multimodal large language models (MLLMs), concerns regarding their security have increasingly captured the attention of both academia and industry. Although MLLMs are vulnerable to jailbreak attacks, designing effective jailbreak attacks poses unique challenges, especially given the highly constrained adversarial capabilities in real-world deployment scenarios. Previous works concentrate risks into a single modality, resulting in limited jailbreak performance. In this paper, we propose a heuristic-induced multimodal risk distribution jailbreak attack method, calle"},"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":"2412.05934","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-12-08T13:20:45Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"7559fdda6e6f85c0387b5a70aa9275c97173742a4f2f65ce7a4518653a94d178","abstract_canon_sha256":"ed463aa3bf19a154ade6647c4548f0efbe415f09076f43d64aa69b7d0f2aeb6a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:55:15.717276Z","signature_b64":"Ce+pT98OnNMniK6JSzcSA0bmQyFep08VxrT4oRsNSQpWxLrNUzx8fsMh2icdIuWk/0eDX0c0aueDUWntJ3vuAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"621472a50cc24b9f5eb6a4686c6a349a689c0f9eac49be6ae0cf8c2d3f48ced8","last_reissued_at":"2026-07-05T11:55:15.716776Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:55:15.716776Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Heuristic-Induced Multimodal Risk Distribution Jailbreak Attack for Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Chu Zhixuan, Duan Ranjie, Huang Yihao, Jia Xiaojun, Jia Xiaoshuang, Li Xinfeng, Ma Teng, Ren Wenqi","submitted_at":"2024-12-08T13:20:45Z","abstract_excerpt":"With the rapid advancement of multimodal large language models (MLLMs), concerns regarding their security have increasingly captured the attention of both academia and industry. Although MLLMs are vulnerable to jailbreak attacks, designing effective jailbreak attacks poses unique challenges, especially given the highly constrained adversarial capabilities in real-world deployment scenarios. Previous works concentrate risks into a single modality, resulting in limited jailbreak performance. In this paper, we propose a heuristic-induced multimodal risk distribution jailbreak attack method, calle"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.05934","kind":"arxiv","version":3},"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/2412.05934/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":"2412.05934","created_at":"2026-07-05T11:55:15.716835+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.05934v3","created_at":"2026-07-05T11:55:15.716835+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.05934","created_at":"2026-07-05T11:55:15.716835+00:00"},{"alias_kind":"pith_short_12","alias_value":"MIKHFJIMYJFZ","created_at":"2026-07-05T11:55:15.716835+00:00"},{"alias_kind":"pith_short_16","alias_value":"MIKHFJIMYJFZ6XVW","created_at":"2026-07-05T11:55:15.716835+00:00"},{"alias_kind":"pith_short_8","alias_value":"MIKHFJIM","created_at":"2026-07-05T11:55:15.716835+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.01837","citing_title":"Benign Inputs, Harmful Outputs: Cross-Modal Jailbreaking via Distributed Semantic Recomposition","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2505.11548","citing_title":"One Shot Dominance: Knowledge Poisoning Attack on Retrieval-Augmented Generation Systems","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2507.21540","citing_title":"PRISM: Programmatic Reasoning with Image Sequence Manipulation for LVLM Jailbreaking","ref_index":34,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ","json":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ.json","graph_json":"https://pith.science/api/pith-number/MIKHFJIMYJFZ6XVWURUGY2RUTJ/graph.json","events_json":"https://pith.science/api/pith-number/MIKHFJIMYJFZ6XVWURUGY2RUTJ/events.json","paper":"https://pith.science/paper/MIKHFJIM"},"agent_actions":{"view_html":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ","download_json":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ.json","view_paper":"https://pith.science/paper/MIKHFJIM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.05934&json=true","fetch_graph":"https://pith.science/api/pith-number/MIKHFJIMYJFZ6XVWURUGY2RUTJ/graph.json","fetch_events":"https://pith.science/api/pith-number/MIKHFJIMYJFZ6XVWURUGY2RUTJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ/action/storage_attestation","attest_author":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ/action/author_attestation","sign_citation":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ/action/citation_signature","submit_replication":"https://pith.science/pith/MIKHFJIMYJFZ6XVWURUGY2RUTJ/action/replication_record"}},"created_at":"2026-07-05T11:55:15.716835+00:00","updated_at":"2026-07-05T11:55:15.716835+00:00"}