{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RJI3KLKZNFXTMBYHN627423OO3","short_pith_number":"pith:RJI3KLKZ","schema_version":"1.0","canonical_sha256":"8a51b52d59696f3607076fb5fe6b6e76e1283d4c02fa962fbce21d6a69a1b473","source":{"kind":"arxiv","id":"2410.21083","version":2},"attestation_state":"computed","paper":{"title":"Stealthy Jailbreak Attacks on Large Language Models via Benign Data Mirroring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Han He, Honglin Mu, Libo Qin, Qingfu Zhu, Qi Shi, Wanxiang Che, Xiaoming Shi, Xudong Han, Yang Xu, Yunlong Feng, Yuxin Zhou, Zeming Liu","submitted_at":"2024-10-28T14:48:05Z","abstract_excerpt":"Large language model (LLM) safety is a critical issue, with numerous studies employing red team testing to enhance model security. Among these, jailbreak methods explore potential vulnerabilities by crafting malicious prompts that induce model outputs contrary to safety alignments. Existing black-box jailbreak methods often rely on model feedback, repeatedly submitting queries with detectable malicious instructions during the attack search process. Although these approaches are effective, the attacks may be intercepted by content moderators during the search process. We propose an improved tra"},"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":"2410.21083","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-10-28T14:48:05Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9fc0918c9bd49c1a75f918df00b6817d6f6523499b828b67af8fbee723053848","abstract_canon_sha256":"0a643221ac6b856c640006fcc70f394a2f6b54513dc2f40dfd3592e956518b38"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:25:12.810554Z","signature_b64":"fUlyp3oVjsy0Z4SETOFfqRGsSYWR0v9mgp1OfOADOUSKErf9OWLB7cTT+R67yFNFKVV5gh/vI1nEV1LK9l7zBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8a51b52d59696f3607076fb5fe6b6e76e1283d4c02fa962fbce21d6a69a1b473","last_reissued_at":"2026-07-05T10:25:12.809508Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:25:12.809508Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Stealthy Jailbreak Attacks on Large Language Models via Benign Data Mirroring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Han He, Honglin Mu, Libo Qin, Qingfu Zhu, Qi Shi, Wanxiang Che, Xiaoming Shi, Xudong Han, Yang Xu, Yunlong Feng, Yuxin Zhou, Zeming Liu","submitted_at":"2024-10-28T14:48:05Z","abstract_excerpt":"Large language model (LLM) safety is a critical issue, with numerous studies employing red team testing to enhance model security. Among these, jailbreak methods explore potential vulnerabilities by crafting malicious prompts that induce model outputs contrary to safety alignments. Existing black-box jailbreak methods often rely on model feedback, repeatedly submitting queries with detectable malicious instructions during the attack search process. Although these approaches are effective, the attacks may be intercepted by content moderators during the search process. We propose an improved tra"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.21083","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/2410.21083/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":"2410.21083","created_at":"2026-07-05T10:25:12.809636+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.21083v2","created_at":"2026-07-05T10:25:12.809636+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.21083","created_at":"2026-07-05T10:25:12.809636+00:00"},{"alias_kind":"pith_short_12","alias_value":"RJI3KLKZNFXT","created_at":"2026-07-05T10:25:12.809636+00:00"},{"alias_kind":"pith_short_16","alias_value":"RJI3KLKZNFXTMBYH","created_at":"2026-07-05T10:25:12.809636+00:00"},{"alias_kind":"pith_short_8","alias_value":"RJI3KLKZ","created_at":"2026-07-05T10:25:12.809636+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2608.10530","citing_title":"On Understanding, Identifying, and Mitigating Vulnerabilities in Agentic Large Language Models","ref_index":89,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3","json":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3.json","graph_json":"https://pith.science/api/pith-number/RJI3KLKZNFXTMBYHN627423OO3/graph.json","events_json":"https://pith.science/api/pith-number/RJI3KLKZNFXTMBYHN627423OO3/events.json","paper":"https://pith.science/paper/RJI3KLKZ"},"agent_actions":{"view_html":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3","download_json":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3.json","view_paper":"https://pith.science/paper/RJI3KLKZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.21083&json=true","fetch_graph":"https://pith.science/api/pith-number/RJI3KLKZNFXTMBYHN627423OO3/graph.json","fetch_events":"https://pith.science/api/pith-number/RJI3KLKZNFXTMBYHN627423OO3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3/action/storage_attestation","attest_author":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3/action/author_attestation","sign_citation":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3/action/citation_signature","submit_replication":"https://pith.science/pith/RJI3KLKZNFXTMBYHN627423OO3/action/replication_record"}},"created_at":"2026-07-05T10:25:12.809636+00:00","updated_at":"2026-07-05T10:25:12.809636+00:00"}