{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:AII7RK7IMEGTSXZ5GUK4ESFSYJ","short_pith_number":"pith:AII7RK7I","canonical_record":{"source":{"id":"2403.08424","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-03-13T11:16:43Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"eea0569776142afcec339f29087bc6003f4dc746f4f27c1f9c590b0492f4e093","abstract_canon_sha256":"eef276a45d69ee2d2102e4d31eb95a8ec1e384ca816dc9bd8e1d15b88a8ae054"},"schema_version":"1.0"},"canonical_sha256":"0211f8abe8610d395f3d3515c248b2c27eb52602ce0bc32bbc228b1d8e859a83","source":{"kind":"arxiv","id":"2403.08424","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.08424","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"arxiv_version","alias_value":"2403.08424v2","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.08424","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_12","alias_value":"AII7RK7IMEGT","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_16","alias_value":"AII7RK7IMEGTSXZ5","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_8","alias_value":"AII7RK7I","created_at":"2026-07-05T09:13:35Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:AII7RK7IMEGTSXZ5GUK4ESFSYJ","target":"record","payload":{"canonical_record":{"source":{"id":"2403.08424","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-03-13T11:16:43Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"eea0569776142afcec339f29087bc6003f4dc746f4f27c1f9c590b0492f4e093","abstract_canon_sha256":"eef276a45d69ee2d2102e4d31eb95a8ec1e384ca816dc9bd8e1d15b88a8ae054"},"schema_version":"1.0"},"canonical_sha256":"0211f8abe8610d395f3d3515c248b2c27eb52602ce0bc32bbc228b1d8e859a83","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:13:35.431021Z","signature_b64":"PtNgBG9vta2JZfpia/S/wMhXExKllJ03nFEmJSbVsVL9ueDYTc9+qmgHs6edR5O/XtFPxAKj+78qbZ1Z3OyQBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0211f8abe8610d395f3d3515c248b2c27eb52602ce0bc32bbc228b1d8e859a83","last_reissued_at":"2026-07-05T09:13:35.430549Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:13:35.430549Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2403.08424","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":"2026-07-05T09:13:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xjrrNGYia6H25f+c/e7gQklVUewXU5nho7akhsMfmDfPe6k/3DogoB09rBEotPE7Tb4YPQoGOwnfbk1ozMQ7BA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T02:52:57.457886Z"},"content_sha256":"f5732c54113fd71ac6dd1697138e896bb4c77639b22908f560c693dfcdd9ef01","schema_version":"1.0","event_id":"sha256:f5732c54113fd71ac6dd1697138e896bb4c77639b22908f560c693dfcdd9ef01"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:AII7RK7IMEGTSXZ5GUK4ESFSYJ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Distract Large Language Models for Automatic Jailbreak Attack","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Guanhua Chen, Yan Yang, Yun Chen, Zeguan Xiao","submitted_at":"2024-03-13T11:16:43Z","abstract_excerpt":"Extensive efforts have been made before the public release of Large language models (LLMs) to align their behaviors with human values. However, even meticulously aligned LLMs remain vulnerable to malicious manipulations such as jailbreaking, leading to unintended behaviors. In this work, we propose a novel black-box jailbreak framework for automated red teaming of LLMs. We designed malicious content concealing and memory reframing with an iterative optimization algorithm to jailbreak LLMs, motivated by the research about the distractibility and over-confidence phenomenon of LLMs. Extensive exp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.08424","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/2403.08424/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":"2026-07-05T09:13:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"j0o1cxBL3iDGSICVTW9Tl9ORxwCHA76FEK1XhZmNoDyYbYeKkIqI4nNQDCUqbjSDLGgBWj88HbQSmbdlQdI+Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T02:52:57.458692Z"},"content_sha256":"7d426a11e641209282133816caed23afadc4b4721e45e6420dcc957cf9d76034","schema_version":"1.0","event_id":"sha256:7d426a11e641209282133816caed23afadc4b4721e45e6420dcc957cf9d76034"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/bundle.json","state_url":"https://pith.science/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/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":"2026-08-10T02:52:57Z","links":{"resolver":"https://pith.science/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ","bundle":"https://pith.science/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/bundle.json","state":"https://pith.science/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AII7RK7IMEGTSXZ5GUK4ESFSYJ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:AII7RK7IMEGTSXZ5GUK4ESFSYJ","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":"eef276a45d69ee2d2102e4d31eb95a8ec1e384ca816dc9bd8e1d15b88a8ae054","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-03-13T11:16:43Z","title_canon_sha256":"eea0569776142afcec339f29087bc6003f4dc746f4f27c1f9c590b0492f4e093"},"schema_version":"1.0","source":{"id":"2403.08424","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2403.08424","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"arxiv_version","alias_value":"2403.08424v2","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2403.08424","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_12","alias_value":"AII7RK7IMEGT","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_16","alias_value":"AII7RK7IMEGTSXZ5","created_at":"2026-07-05T09:13:35Z"},{"alias_kind":"pith_short_8","alias_value":"AII7RK7I","created_at":"2026-07-05T09:13:35Z"}],"graph_snapshots":[{"event_id":"sha256:7d426a11e641209282133816caed23afadc4b4721e45e6420dcc957cf9d76034","target":"graph","created_at":"2026-07-05T09:13:35Z","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/2403.08424/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Extensive efforts have been made before the public release of Large language models (LLMs) to align their behaviors with human values. However, even meticulously aligned LLMs remain vulnerable to malicious manipulations such as jailbreaking, leading to unintended behaviors. In this work, we propose a novel black-box jailbreak framework for automated red teaming of LLMs. We designed malicious content concealing and memory reframing with an iterative optimization algorithm to jailbreak LLMs, motivated by the research about the distractibility and over-confidence phenomenon of LLMs. Extensive exp","authors_text":"Guanhua Chen, Yan Yang, Yun Chen, Zeguan Xiao","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-03-13T11:16:43Z","title":"Distract Large Language Models for Automatic Jailbreak Attack"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2403.08424","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:f5732c54113fd71ac6dd1697138e896bb4c77639b22908f560c693dfcdd9ef01","target":"record","created_at":"2026-07-05T09:13:35Z","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":"eef276a45d69ee2d2102e4d31eb95a8ec1e384ca816dc9bd8e1d15b88a8ae054","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-03-13T11:16:43Z","title_canon_sha256":"eea0569776142afcec339f29087bc6003f4dc746f4f27c1f9c590b0492f4e093"},"schema_version":"1.0","source":{"id":"2403.08424","kind":"arxiv","version":2}},"canonical_sha256":"0211f8abe8610d395f3d3515c248b2c27eb52602ce0bc32bbc228b1d8e859a83","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0211f8abe8610d395f3d3515c248b2c27eb52602ce0bc32bbc228b1d8e859a83","first_computed_at":"2026-07-05T09:13:35.430549Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:13:35.430549Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"PtNgBG9vta2JZfpia/S/wMhXExKllJ03nFEmJSbVsVL9ueDYTc9+qmgHs6edR5O/XtFPxAKj+78qbZ1Z3OyQBw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:13:35.431021Z","signed_message":"canonical_sha256_bytes"},"source_id":"2403.08424","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f5732c54113fd71ac6dd1697138e896bb4c77639b22908f560c693dfcdd9ef01","sha256:7d426a11e641209282133816caed23afadc4b4721e45e6420dcc957cf9d76034"],"state_sha256":"a1636a66f898965814f31e103e0c73088ca90a1341d69e65af1ff344a2da3804"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"u1TgvEPuu+1F8uYwZqcFIK8jLIVTFWUw9h5fzX4rYoJwWJD2edrS833fgdrqRORmhFikhQQeCywSykMYJQVkAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T02:52:57.466313Z","bundle_sha256":"3c89001ea2644e879448b36fdce636ee18d0bab397d88e82fc5f6b4a54c08386"}}