{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:64ZWKEDS4VIUSNDTPANX4GZ52Y","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":"e6dd431c296cc89dde9fc3b6999d91d6d6779294929ba66d72899b7baf0fdb7e","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T16:13:24Z","title_canon_sha256":"e95d45ac08d7e5fb69a43e029147f02c6d2c93a5a644823619bb14ad52459645"},"schema_version":"1.0","source":{"id":"2401.06373","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.06373","created_at":"2026-07-05T07:36:57Z"},{"alias_kind":"arxiv_version","alias_value":"2401.06373v2","created_at":"2026-07-05T07:36:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.06373","created_at":"2026-07-05T07:36:57Z"},{"alias_kind":"pith_short_12","alias_value":"64ZWKEDS4VIU","created_at":"2026-07-05T07:36:57Z"},{"alias_kind":"pith_short_16","alias_value":"64ZWKEDS4VIUSNDT","created_at":"2026-07-05T07:36:57Z"},{"alias_kind":"pith_short_8","alias_value":"64ZWKEDS","created_at":"2026-07-05T07:36:57Z"}],"graph_snapshots":[{"event_id":"sha256:856956a1718ab302c6a5d5fb220fb1531f9c96e3680de1ce7ac1e66e76770f53","target":"graph","created_at":"2026-07-05T07:36:57Z","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/2401.06373/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Most traditional AI safety research has approached AI models as machines and centered on algorithm-focused attacks developed by security experts. As large language models (LLMs) become increasingly common and competent, non-expert users can also impose risks during daily interactions. This paper introduces a new perspective to jailbreak LLMs as human-like communicators, to explore this overlooked intersection between everyday language interaction and AI safety. Specifically, we study how to persuade LLMs to jailbreak them. First, we propose a persuasion taxonomy derived from decades of social ","authors_text":"Diyi Yang, Hongpeng Lin, Jingwen Zhang, Ruoxi Jia, Weiyan Shi, Yi Zeng","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T16:13:24Z","title":"How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.06373","kind":"arxiv","version":2},"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":"sha256:9fee07785e1f7808b5c304e46290f472fda2cab561bc583bbd67806ff4589d67","target":"record","created_at":"2026-07-05T07:36:57Z","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":"e6dd431c296cc89dde9fc3b6999d91d6d6779294929ba66d72899b7baf0fdb7e","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-01-12T16:13:24Z","title_canon_sha256":"e95d45ac08d7e5fb69a43e029147f02c6d2c93a5a644823619bb14ad52459645"},"schema_version":"1.0","source":{"id":"2401.06373","kind":"arxiv","version":2}},"canonical_sha256":"f733651072e551493473781b7e1b3dd601855f714a4affefc680e58bc40291c3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f733651072e551493473781b7e1b3dd601855f714a4affefc680e58bc40291c3","first_computed_at":"2026-07-05T07:36:57.295780Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:36:57.295780Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"imCD/VC6sjHP1Qmhl5E5Bm9+TKYCz72boanFOEjrIC1sLGSgDlONPvHy1bPU1DTmwStJSZrr8/pn/vGj97KOBg==","signature_status":"signed_v1","signed_at":"2026-07-05T07:36:57.296209Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.06373","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9fee07785e1f7808b5c304e46290f472fda2cab561bc583bbd67806ff4589d67","sha256:856956a1718ab302c6a5d5fb220fb1531f9c96e3680de1ce7ac1e66e76770f53"],"state_sha256":"a6bfd1be8a83aadd602552d0eb3fbf90793f3c430a8c255559eaa58edd34a330"}