{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:MDDSRA4ZDK4OCZOBSKZVLOCX42","short_pith_number":"pith:MDDSRA4Z","canonical_record":{"source":{"id":"2311.08268","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T16:02:16Z","cross_cats_sorted":[],"title_canon_sha256":"2d3769c3953470a7661016b75c62eb8b2746967d8d399a5012ef3f9fcd2a3d70","abstract_canon_sha256":"a66302d8c7c10a3b16a18af457054cbc9dbba19cd65699d235d78e97c7c6affd"},"schema_version":"1.0"},"canonical_sha256":"60c72883991ab8e165c192b355b857e69b8dfc8fc21b4b8f04115823a1faad74","source":{"kind":"arxiv","id":"2311.08268","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.08268","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"arxiv_version","alias_value":"2311.08268v4","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.08268","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_12","alias_value":"MDDSRA4ZDK4O","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_16","alias_value":"MDDSRA4ZDK4OCZOB","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_8","alias_value":"MDDSRA4Z","created_at":"2026-07-05T08:05:05Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:MDDSRA4ZDK4OCZOBSKZVLOCX42","target":"record","payload":{"canonical_record":{"source":{"id":"2311.08268","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T16:02:16Z","cross_cats_sorted":[],"title_canon_sha256":"2d3769c3953470a7661016b75c62eb8b2746967d8d399a5012ef3f9fcd2a3d70","abstract_canon_sha256":"a66302d8c7c10a3b16a18af457054cbc9dbba19cd65699d235d78e97c7c6affd"},"schema_version":"1.0"},"canonical_sha256":"60c72883991ab8e165c192b355b857e69b8dfc8fc21b4b8f04115823a1faad74","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:05:05.058249Z","signature_b64":"IcIGoT5NTUIn+QZUiWUUuC8vJ0CssW9PoWTosAT9HtHRArSDc0xnCxEazMxtnAdEE0tyHRqVnMnLpvr1zClRAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"60c72883991ab8e165c192b355b857e69b8dfc8fc21b4b8f04115823a1faad74","last_reissued_at":"2026-07-05T08:05:05.057777Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:05:05.057777Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2311.08268","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:05:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rXSPPNsXsdWAHWdF7zd10sAUKPrutGm0PRjG5uMkxc//dYaqo/bnBQKoUp+OvIkTiIj4oupE3+xBNd1V5VZABw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T23:30:38.329001Z"},"content_sha256":"6d1e30f2497a01d3fe30aa655345644c15d971863c117075221228f8c0163339","schema_version":"1.0","event_id":"sha256:6d1e30f2497a01d3fe30aa655345644c15d971863c117075221228f8c0163339"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:MDDSRA4ZDK4OCZOBSKZVLOCX42","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Wolf in Sheep's Clothing: Generalized Nested Jailbreak Prompts can Fool Large Language Models Easily","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Dan Ma, Jiajun Chen, Jun Kuang, Peng Ding, Shujian Huang, Xuezhi Cao, Yunsen Xian","submitted_at":"2023-11-14T16:02:16Z","abstract_excerpt":"Large Language Models (LLMs), such as ChatGPT and GPT-4, are designed to provide useful and safe responses. However, adversarial prompts known as 'jailbreaks' can circumvent safeguards, leading LLMs to generate potentially harmful content. Exploring jailbreak prompts can help to better reveal the weaknesses of LLMs and further steer us to secure them. Unfortunately, existing jailbreak methods either suffer from intricate manual design or require optimization on other white-box models, which compromises either generalization or efficiency. In this paper, we generalize jailbreak prompt attacks i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.08268","kind":"arxiv","version":4},"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/2311.08268/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:05:05Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"o5B6fq88IZ3ytqk7emT78IgxkqQi3mvTuiqhQeIm17rQ/Z1vA7kvmONHJvRzClrk5BtO2js56WUc3PJEECelDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-06T23:30:38.329392Z"},"content_sha256":"75160c9ca748310fa2d2588b6d042aa00e7534ed14443c770ace2b78e670c247","schema_version":"1.0","event_id":"sha256:75160c9ca748310fa2d2588b6d042aa00e7534ed14443c770ace2b78e670c247"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/bundle.json","state_url":"https://pith.science/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-07-06T23:30:38Z","links":{"resolver":"https://pith.science/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42","bundle":"https://pith.science/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/bundle.json","state":"https://pith.science/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/state.json","well_known_bundle":"https://pith.science/.well-known/pith/MDDSRA4ZDK4OCZOBSKZVLOCX42/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:MDDSRA4ZDK4OCZOBSKZVLOCX42","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":"a66302d8c7c10a3b16a18af457054cbc9dbba19cd65699d235d78e97c7c6affd","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T16:02:16Z","title_canon_sha256":"2d3769c3953470a7661016b75c62eb8b2746967d8d399a5012ef3f9fcd2a3d70"},"schema_version":"1.0","source":{"id":"2311.08268","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2311.08268","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"arxiv_version","alias_value":"2311.08268v4","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.08268","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_12","alias_value":"MDDSRA4ZDK4O","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_16","alias_value":"MDDSRA4ZDK4OCZOB","created_at":"2026-07-05T08:05:05Z"},{"alias_kind":"pith_short_8","alias_value":"MDDSRA4Z","created_at":"2026-07-05T08:05:05Z"}],"graph_snapshots":[{"event_id":"sha256:75160c9ca748310fa2d2588b6d042aa00e7534ed14443c770ace2b78e670c247","target":"graph","created_at":"2026-07-05T08:05:05Z","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/2311.08268/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs), such as ChatGPT and GPT-4, are designed to provide useful and safe responses. However, adversarial prompts known as 'jailbreaks' can circumvent safeguards, leading LLMs to generate potentially harmful content. Exploring jailbreak prompts can help to better reveal the weaknesses of LLMs and further steer us to secure them. Unfortunately, existing jailbreak methods either suffer from intricate manual design or require optimization on other white-box models, which compromises either generalization or efficiency. In this paper, we generalize jailbreak prompt attacks i","authors_text":"Dan Ma, Jiajun Chen, Jun Kuang, Peng Ding, Shujian Huang, Xuezhi Cao, Yunsen Xian","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T16:02:16Z","title":"A Wolf in Sheep's Clothing: Generalized Nested Jailbreak Prompts can Fool Large Language Models Easily"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.08268","kind":"arxiv","version":4},"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":"sha256:6d1e30f2497a01d3fe30aa655345644c15d971863c117075221228f8c0163339","target":"record","created_at":"2026-07-05T08:05:05Z","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":"a66302d8c7c10a3b16a18af457054cbc9dbba19cd65699d235d78e97c7c6affd","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T16:02:16Z","title_canon_sha256":"2d3769c3953470a7661016b75c62eb8b2746967d8d399a5012ef3f9fcd2a3d70"},"schema_version":"1.0","source":{"id":"2311.08268","kind":"arxiv","version":4}},"canonical_sha256":"60c72883991ab8e165c192b355b857e69b8dfc8fc21b4b8f04115823a1faad74","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"60c72883991ab8e165c192b355b857e69b8dfc8fc21b4b8f04115823a1faad74","first_computed_at":"2026-07-05T08:05:05.057777Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:05:05.057777Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"IcIGoT5NTUIn+QZUiWUUuC8vJ0CssW9PoWTosAT9HtHRArSDc0xnCxEazMxtnAdEE0tyHRqVnMnLpvr1zClRAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:05:05.058249Z","signed_message":"canonical_sha256_bytes"},"source_id":"2311.08268","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6d1e30f2497a01d3fe30aa655345644c15d971863c117075221228f8c0163339","sha256:75160c9ca748310fa2d2588b6d042aa00e7534ed14443c770ace2b78e670c247"],"state_sha256":"cccf925ec102e1c2684821e9bafb17971f924fabf7077c6948a8e3b2d24d8902"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZsMjx4E67ySOIKbSa5xZ2NhmCRXbWN5FNwVpR7hpKtIJ2HF/PU2EOjpcr9IvGDTNgMAGXEpiZpkHKay8KQ85AQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-06T23:30:38.331381Z","bundle_sha256":"941c219af49a0bbedf70f172b3f32869434f873155e21b217b1c31085d15d714"}}