{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:TIY7QE3PDH5GQJZLZ6MZCTA4OB","short_pith_number":"pith:TIY7QE3P","canonical_record":{"source":{"id":"2406.11668","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T15:51:01Z","cross_cats_sorted":[],"title_canon_sha256":"f3d2388261b0800d3f1aa17aba1751a581a4d704c3b007b0b7b1541896567f32","abstract_canon_sha256":"8151e54962e731d33193da9a07ce5712177c9e515d84e1013c383fa2e0d0e945"},"schema_version":"1.0"},"canonical_sha256":"9a31f8136f19fa68272bcf99914c1c70705b28cc648740d63ee09fee14fa4fbe","source":{"kind":"arxiv","id":"2406.11668","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.11668","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"arxiv_version","alias_value":"2406.11668v2","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11668","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_12","alias_value":"TIY7QE3PDH5G","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_16","alias_value":"TIY7QE3PDH5GQJZL","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_8","alias_value":"TIY7QE3P","created_at":"2026-07-05T10:08:18Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:TIY7QE3PDH5GQJZLZ6MZCTA4OB","target":"record","payload":{"canonical_record":{"source":{"id":"2406.11668","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T15:51:01Z","cross_cats_sorted":[],"title_canon_sha256":"f3d2388261b0800d3f1aa17aba1751a581a4d704c3b007b0b7b1541896567f32","abstract_canon_sha256":"8151e54962e731d33193da9a07ce5712177c9e515d84e1013c383fa2e0d0e945"},"schema_version":"1.0"},"canonical_sha256":"9a31f8136f19fa68272bcf99914c1c70705b28cc648740d63ee09fee14fa4fbe","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:08:18.372971Z","signature_b64":"/gSHk5sj1aDlToYWcFvsKFrzNDCwFH8Nna/gVX3ifiPRqK2RLfntz8/UTHt9PVAXam4sQjMh1KnSnBe6zmn0DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9a31f8136f19fa68272bcf99914c1c70705b28cc648740d63ee09fee14fa4fbe","last_reissued_at":"2026-07-05T10:08:18.372535Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:08:18.372535Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.11668","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:08:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"KXZaVJaXM64D+M9KFSo4M7UwryTFZhXHSlsw1lIDQgG/y3bbwW0jfT1F5OqR8rA/7pWtk2cfgoZ+dN2Sg4xVCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T07:58:52.328179Z"},"content_sha256":"5d6c6259a63c69125bf03c3917be7faca2d2a27d0a76636652ab10378476f4b1","schema_version":"1.0","event_id":"sha256:5d6c6259a63c69125bf03c3917be7faca2d2a27d0a76636652ab10378476f4b1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:TIY7QE3PDH5GQJZLZ6MZCTA4OB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"\"Not Aligned\" is Not \"Malicious\": Being Careful about Hallucinations of Large Language Models' Jailbreak","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Baolong Bi, Jiayi Mao, Lingrui Mei, Shenghua Liu, Xueqi Cheng, Yiwei Wang","submitted_at":"2024-06-17T15:51:01Z","abstract_excerpt":"\"Jailbreak\" is a major safety concern of Large Language Models (LLMs), which occurs when malicious prompts lead LLMs to produce harmful outputs, raising issues about the reliability and safety of LLMs. Therefore, an effective evaluation of jailbreaks is very crucial to develop its mitigation strategies. However, our research reveals that many jailbreaks identified by current evaluations may actually be hallucinations-erroneous outputs that are mistaken for genuine safety breaches. This finding suggests that some perceived vulnerabilities might not represent actual threats, indicating a need fo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11668","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/2406.11668/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":"2026-07-05T10:08:18Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6QG4lecAglwQpMbDQul6++Uryy0WKVBtiKRVpDdhf+n3VYvhZK4SKJ1p67IwkE21iDXtrOAfXjqCh6iX+Us8CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-04T07:58:52.329093Z"},"content_sha256":"342b73fab8c750ad66a05d6a768abe3a7590677247d702ee4c0c8ae80eda69cc","schema_version":"1.0","event_id":"sha256:342b73fab8c750ad66a05d6a768abe3a7590677247d702ee4c0c8ae80eda69cc"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/bundle.json","state_url":"https://pith.science/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/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":"2026-08-04T07:58:52Z","links":{"resolver":"https://pith.science/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB","bundle":"https://pith.science/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/bundle.json","state":"https://pith.science/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TIY7QE3PDH5GQJZLZ6MZCTA4OB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:TIY7QE3PDH5GQJZLZ6MZCTA4OB","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":"8151e54962e731d33193da9a07ce5712177c9e515d84e1013c383fa2e0d0e945","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T15:51:01Z","title_canon_sha256":"f3d2388261b0800d3f1aa17aba1751a581a4d704c3b007b0b7b1541896567f32"},"schema_version":"1.0","source":{"id":"2406.11668","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.11668","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"arxiv_version","alias_value":"2406.11668v2","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.11668","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_12","alias_value":"TIY7QE3PDH5G","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_16","alias_value":"TIY7QE3PDH5GQJZL","created_at":"2026-07-05T10:08:18Z"},{"alias_kind":"pith_short_8","alias_value":"TIY7QE3P","created_at":"2026-07-05T10:08:18Z"}],"graph_snapshots":[{"event_id":"sha256:342b73fab8c750ad66a05d6a768abe3a7590677247d702ee4c0c8ae80eda69cc","target":"graph","created_at":"2026-07-05T10:08:18Z","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/2406.11668/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"\"Jailbreak\" is a major safety concern of Large Language Models (LLMs), which occurs when malicious prompts lead LLMs to produce harmful outputs, raising issues about the reliability and safety of LLMs. Therefore, an effective evaluation of jailbreaks is very crucial to develop its mitigation strategies. However, our research reveals that many jailbreaks identified by current evaluations may actually be hallucinations-erroneous outputs that are mistaken for genuine safety breaches. This finding suggests that some perceived vulnerabilities might not represent actual threats, indicating a need fo","authors_text":"Baolong Bi, Jiayi Mao, Lingrui Mei, Shenghua Liu, Xueqi Cheng, Yiwei Wang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T15:51:01Z","title":"\"Not Aligned\" is Not \"Malicious\": Being Careful about Hallucinations of Large Language Models' Jailbreak"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.11668","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":"sha256:5d6c6259a63c69125bf03c3917be7faca2d2a27d0a76636652ab10378476f4b1","target":"record","created_at":"2026-07-05T10:08:18Z","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":"8151e54962e731d33193da9a07ce5712177c9e515d84e1013c383fa2e0d0e945","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T15:51:01Z","title_canon_sha256":"f3d2388261b0800d3f1aa17aba1751a581a4d704c3b007b0b7b1541896567f32"},"schema_version":"1.0","source":{"id":"2406.11668","kind":"arxiv","version":2}},"canonical_sha256":"9a31f8136f19fa68272bcf99914c1c70705b28cc648740d63ee09fee14fa4fbe","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9a31f8136f19fa68272bcf99914c1c70705b28cc648740d63ee09fee14fa4fbe","first_computed_at":"2026-07-05T10:08:18.372535Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:08:18.372535Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/gSHk5sj1aDlToYWcFvsKFrzNDCwFH8Nna/gVX3ifiPRqK2RLfntz8/UTHt9PVAXam4sQjMh1KnSnBe6zmn0DQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:08:18.372971Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.11668","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:5d6c6259a63c69125bf03c3917be7faca2d2a27d0a76636652ab10378476f4b1","sha256:342b73fab8c750ad66a05d6a768abe3a7590677247d702ee4c0c8ae80eda69cc"],"state_sha256":"c948e79801812b91abce37a31257b401751774cf9ea9e55b58025be1161d7dfb"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"U6nc3/PSMrh0uxuoabgCfLLDliOzUg4Qobsmm248cG5/CPkKGMuYBKFFKa7pHf6beEZi/Pe+LNSt5krWecU5Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-04T07:58:52.379386Z","bundle_sha256":"b8fe1bfd161c914d2d01123d172d834f30521a0ca4eae08909f67ecf862b2558"}}