{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:TQUFZOFVPWOZDSW7V23T5LN5C3","short_pith_number":"pith:TQUFZOFV","schema_version":"1.0","canonical_sha256":"9c285cb8b57d9d91cadfaeb73eadbd16caea4569fa64901bbddc5a127c534172","source":{"kind":"arxiv","id":"2408.04686","version":1},"attestation_state":"computed","paper":{"title":"Multi-Turn Context Jailbreak Attack on Large Language Models From First Principles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Deyue Zhang, Dongdong Yang, Hui Li, Quanchen Zou, Xiongtao Sun","submitted_at":"2024-08-08T09:18:47Z","abstract_excerpt":"Large language models (LLMs) have significantly enhanced the performance of numerous applications, from intelligent conversations to text generation. However, their inherent security vulnerabilities have become an increasingly significant challenge, especially with respect to jailbreak attacks. Attackers can circumvent the security mechanisms of these LLMs, breaching security constraints and causing harmful outputs. Focusing on multi-turn semantic jailbreak attacks, we observe that existing methods lack specific considerations for the role of multiturn dialogues in attack strategies, leading t"},"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":"2408.04686","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-08-08T09:18:47Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"0fb34f4d6aa86a164e9bca9b3b6587625589420400eff2ae70aa6c13d01ada2b","abstract_canon_sha256":"74522533b3b9e1e64ae3e947f5f86b09147f5aee4ee7f7b2f4e95370317fdc53"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:53:42.422982Z","signature_b64":"8Ch3s7OQeUSoHMvtAM8+2122/pvtiYNf3qPQi33ggFABDYY+9X96KmWlYWUdjEgGtCIHwrZId8mM5UgV3224AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9c285cb8b57d9d91cadfaeb73eadbd16caea4569fa64901bbddc5a127c534172","last_reissued_at":"2026-07-05T08:53:42.422511Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:53:42.422511Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Turn Context Jailbreak Attack on Large Language Models From First Principles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Deyue Zhang, Dongdong Yang, Hui Li, Quanchen Zou, Xiongtao Sun","submitted_at":"2024-08-08T09:18:47Z","abstract_excerpt":"Large language models (LLMs) have significantly enhanced the performance of numerous applications, from intelligent conversations to text generation. However, their inherent security vulnerabilities have become an increasingly significant challenge, especially with respect to jailbreak attacks. Attackers can circumvent the security mechanisms of these LLMs, breaching security constraints and causing harmful outputs. Focusing on multi-turn semantic jailbreak attacks, we observe that existing methods lack specific considerations for the role of multiturn dialogues in attack strategies, leading t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.04686","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/2408.04686/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":"2408.04686","created_at":"2026-07-05T08:53:42.422571+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.04686v1","created_at":"2026-07-05T08:53:42.422571+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.04686","created_at":"2026-07-05T08:53:42.422571+00:00"},{"alias_kind":"pith_short_12","alias_value":"TQUFZOFVPWOZ","created_at":"2026-07-05T08:53:42.422571+00:00"},{"alias_kind":"pith_short_16","alias_value":"TQUFZOFVPWOZDSW7","created_at":"2026-07-05T08:53:42.422571+00:00"},{"alias_kind":"pith_short_8","alias_value":"TQUFZOFV","created_at":"2026-07-05T08:53:42.422571+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25476","citing_title":"A Red Teaming Framework for Large Language Models: A Case Study on Faithfulness Evaluation","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20408","citing_title":"NRT-Bench: Benchmarking Multi-Turn Red-Teaming of LLM Operator Agents in Safety-Critical Control Rooms","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00974","citing_title":"SRTJ: Self-Evolving Rule-Driven Training-Free LLM Jailbreaking","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11309","citing_title":"The Salami Slicing Threat: Exploiting Cumulative Risks in LLM Systems","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3","json":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3.json","graph_json":"https://pith.science/api/pith-number/TQUFZOFVPWOZDSW7V23T5LN5C3/graph.json","events_json":"https://pith.science/api/pith-number/TQUFZOFVPWOZDSW7V23T5LN5C3/events.json","paper":"https://pith.science/paper/TQUFZOFV"},"agent_actions":{"view_html":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3","download_json":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3.json","view_paper":"https://pith.science/paper/TQUFZOFV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.04686&json=true","fetch_graph":"https://pith.science/api/pith-number/TQUFZOFVPWOZDSW7V23T5LN5C3/graph.json","fetch_events":"https://pith.science/api/pith-number/TQUFZOFVPWOZDSW7V23T5LN5C3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3/action/storage_attestation","attest_author":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3/action/author_attestation","sign_citation":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3/action/citation_signature","submit_replication":"https://pith.science/pith/TQUFZOFVPWOZDSW7V23T5LN5C3/action/replication_record"}},"created_at":"2026-07-05T08:53:42.422571+00:00","updated_at":"2026-07-05T08:53:42.422571+00:00"}