{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:LYJH7TDII7HHDXDSMDO4AVAYDZ","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"8b61efadb74177eaddc56150b6f24e03c330df9e67e6ed78541eab2c495341cc","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-18T09:28:58Z","title_canon_sha256":"3a959cf138fe1c20c3c9a5f85df73441983f07ceac8fc8f76af031b24e480c93"},"schema_version":"1.0","source":{"id":"2411.11407","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2411.11407","created_at":"2026-07-05T09:36:55Z"},{"alias_kind":"arxiv_version","alias_value":"2411.11407v1","created_at":"2026-07-05T09:36:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.11407","created_at":"2026-07-05T09:36:55Z"},{"alias_kind":"pith_short_12","alias_value":"LYJH7TDII7HH","created_at":"2026-07-05T09:36:55Z"},{"alias_kind":"pith_short_16","alias_value":"LYJH7TDII7HHDXDS","created_at":"2026-07-05T09:36:55Z"},{"alias_kind":"pith_short_8","alias_value":"LYJH7TDI","created_at":"2026-07-05T09:36:55Z"}],"graph_snapshots":[{"event_id":"sha256:25ed33bfbd87e3c6d2c7331bad38c106c63c579052167d6347dd7406866f44f2","target":"graph","created_at":"2026-07-05T09:36:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2411.11407/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The widespread deployment of large language models (LLMs) across various domains has showcased their immense potential while exposing significant safety vulnerabilities. A major concern is ensuring that LLM-generated content aligns with human values. Existing jailbreak techniques reveal how this alignment can be compromised through specific prompts or adversarial suffixes. In this study, we introduce a new threat: LLMs' bias toward authority. While this inherent bias can improve the quality of outputs generated by LLMs, it also introduces a potential vulnerability, increasing the risk of produ","authors_text":"Jizhong Han, Songlin Hu, Xikang Yang, Xuehai Tang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-18T09:28:58Z","title":"The Dark Side of Trust: Authority Citation-Driven Jailbreak Attacks on Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.11407","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:725b7dcd0a0fbe0947048574fc25e50d6964a62237344b33e05411cd24963b4c","target":"record","created_at":"2026-07-05T09:36:55Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"8b61efadb74177eaddc56150b6f24e03c330df9e67e6ed78541eab2c495341cc","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-11-18T09:28:58Z","title_canon_sha256":"3a959cf138fe1c20c3c9a5f85df73441983f07ceac8fc8f76af031b24e480c93"},"schema_version":"1.0","source":{"id":"2411.11407","kind":"arxiv","version":1}},"canonical_sha256":"5e127fcc6847ce71dc7260ddc054181e7a0838b20c6e0c7320d522f595b34080","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5e127fcc6847ce71dc7260ddc054181e7a0838b20c6e0c7320d522f595b34080","first_computed_at":"2026-07-05T09:36:55.903168Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:36:55.903168Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"aUbA61QTiPzkOjGjGCXO20Z0/4eZs4SlGHorg92z2XYAtczGtUGenHnxrxaG3AL8AO4xqixxJ7qcTke3QbP3BA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:36:55.903624Z","signed_message":"canonical_sha256_bytes"},"source_id":"2411.11407","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:725b7dcd0a0fbe0947048574fc25e50d6964a62237344b33e05411cd24963b4c","sha256:25ed33bfbd87e3c6d2c7331bad38c106c63c579052167d6347dd7406866f44f2"],"state_sha256":"b4e00afb073bf421b89356d3af3cace2ac5067bef6c104234ce79da5eede8e3b"}