{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:TVC6GVW7DXGN2G2KR47KHKWSVL","short_pith_number":"pith:TVC6GVW7","canonical_record":{"source":{"id":"2405.09113","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-15T06:11:24Z","cross_cats_sorted":[],"title_canon_sha256":"886c1b9dba175745cbe294d2893cc351ec34fe0c5a06dbdb70e17267f0c6e660","abstract_canon_sha256":"0f39411e9d536897329e0049ee43345a4678310e49bb61818304dd636d0e8d01"},"schema_version":"1.0"},"canonical_sha256":"9d45e356df1dccdd1b4a8f3ea3aad2aaf6188a62f88cf1c5f79fc356c96e6382","source":{"kind":"arxiv","id":"2405.09113","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.09113","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"arxiv_version","alias_value":"2405.09113v2","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.09113","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_12","alias_value":"TVC6GVW7DXGN","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_16","alias_value":"TVC6GVW7DXGN2G2K","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_8","alias_value":"TVC6GVW7","created_at":"2026-07-05T10:13:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:TVC6GVW7DXGN2G2KR47KHKWSVL","target":"record","payload":{"canonical_record":{"source":{"id":"2405.09113","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-15T06:11:24Z","cross_cats_sorted":[],"title_canon_sha256":"886c1b9dba175745cbe294d2893cc351ec34fe0c5a06dbdb70e17267f0c6e660","abstract_canon_sha256":"0f39411e9d536897329e0049ee43345a4678310e49bb61818304dd636d0e8d01"},"schema_version":"1.0"},"canonical_sha256":"9d45e356df1dccdd1b4a8f3ea3aad2aaf6188a62f88cf1c5f79fc356c96e6382","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:13:00.428061Z","signature_b64":"XCp4uyRT5ezRJLCEAR5qfIinqotxj4T98zn6xtzSoMi+ALtyP08CI9X5/Qj2nNDNEBiysNi++B7YfSmgBUgSBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9d45e356df1dccdd1b4a8f3ea3aad2aaf6188a62f88cf1c5f79fc356c96e6382","last_reissued_at":"2026-07-05T10:13:00.427558Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:13:00.427558Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.09113","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":"2026-07-05T10:13:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"gJBPfhg9nrndVeUt/E1ir8P2TPRueckeIqL50bPinw2veuy/zVMi7CiAYcDVlha01a18CimcbcIAsRfYAaZBDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T10:03:34.852104Z"},"content_sha256":"041b22c07e1a17a1e24cad5d8722d275f3bcdd100afcf6d33e6e96fb23ece29b","schema_version":"1.0","event_id":"sha256:041b22c07e1a17a1e24cad5d8722d275f3bcdd100afcf6d33e6e96fb23ece29b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:TVC6GVW7DXGN2G2KR47KHKWSVL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kai Chen, Kai Hu, Lijun Yu, Matt Fredrikson, Tianjun Yao, Weichen Yu, Wenhe Liu, Xiang Li, Yining Li, Zhiqiang Shen","submitted_at":"2024-05-15T06:11:24Z","abstract_excerpt":"Recent research indicates that large language models (LLMs) are susceptible to jailbreaking attacks that can generate harmful content. This paper introduces a novel token-level attack method, Adaptive Dense-to-Sparse Constrained Optimization (ADC), which has been shown to successfully jailbreak multiple open-source LLMs. Drawing inspiration from the difficulties of discrete token optimization, our method relaxes the discrete jailbreak optimization into a continuous optimization process while gradually increasing the sparsity of the optimizing vectors. This technique effectively bridges the gap"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.09113","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/2405.09113/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":"2026-07-05T10:13:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HdUCpuazhSIzhoWy98EDsssUnS/Rjp034fqq+nbLJ9LPht1gZKVnKa7VwXOVfxdQyv9UfEJJDIBg5OdZGqBkDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T10:03:34.852777Z"},"content_sha256":"403db6a2404a171dbf8777589de74887e7536461d975481aef38f152fbbc6dda","schema_version":"1.0","event_id":"sha256:403db6a2404a171dbf8777589de74887e7536461d975481aef38f152fbbc6dda"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/bundle.json","state_url":"https://pith.science/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/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":"2026-08-06T10:03:34Z","links":{"resolver":"https://pith.science/pith/TVC6GVW7DXGN2G2KR47KHKWSVL","bundle":"https://pith.science/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/bundle.json","state":"https://pith.science/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TVC6GVW7DXGN2G2KR47KHKWSVL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:TVC6GVW7DXGN2G2KR47KHKWSVL","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":"0f39411e9d536897329e0049ee43345a4678310e49bb61818304dd636d0e8d01","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-15T06:11:24Z","title_canon_sha256":"886c1b9dba175745cbe294d2893cc351ec34fe0c5a06dbdb70e17267f0c6e660"},"schema_version":"1.0","source":{"id":"2405.09113","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.09113","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"arxiv_version","alias_value":"2405.09113v2","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.09113","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_12","alias_value":"TVC6GVW7DXGN","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_16","alias_value":"TVC6GVW7DXGN2G2K","created_at":"2026-07-05T10:13:00Z"},{"alias_kind":"pith_short_8","alias_value":"TVC6GVW7","created_at":"2026-07-05T10:13:00Z"}],"graph_snapshots":[{"event_id":"sha256:403db6a2404a171dbf8777589de74887e7536461d975481aef38f152fbbc6dda","target":"graph","created_at":"2026-07-05T10:13:00Z","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/2405.09113/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent research indicates that large language models (LLMs) are susceptible to jailbreaking attacks that can generate harmful content. This paper introduces a novel token-level attack method, Adaptive Dense-to-Sparse Constrained Optimization (ADC), which has been shown to successfully jailbreak multiple open-source LLMs. Drawing inspiration from the difficulties of discrete token optimization, our method relaxes the discrete jailbreak optimization into a continuous optimization process while gradually increasing the sparsity of the optimizing vectors. This technique effectively bridges the gap","authors_text":"Kai Chen, Kai Hu, Lijun Yu, Matt Fredrikson, Tianjun Yao, Weichen Yu, Wenhe Liu, Xiang Li, Yining Li, Zhiqiang Shen","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-15T06:11:24Z","title":"Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.09113","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":"sha256:041b22c07e1a17a1e24cad5d8722d275f3bcdd100afcf6d33e6e96fb23ece29b","target":"record","created_at":"2026-07-05T10:13:00Z","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":"0f39411e9d536897329e0049ee43345a4678310e49bb61818304dd636d0e8d01","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-15T06:11:24Z","title_canon_sha256":"886c1b9dba175745cbe294d2893cc351ec34fe0c5a06dbdb70e17267f0c6e660"},"schema_version":"1.0","source":{"id":"2405.09113","kind":"arxiv","version":2}},"canonical_sha256":"9d45e356df1dccdd1b4a8f3ea3aad2aaf6188a62f88cf1c5f79fc356c96e6382","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9d45e356df1dccdd1b4a8f3ea3aad2aaf6188a62f88cf1c5f79fc356c96e6382","first_computed_at":"2026-07-05T10:13:00.427558Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:13:00.427558Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"XCp4uyRT5ezRJLCEAR5qfIinqotxj4T98zn6xtzSoMi+ALtyP08CI9X5/Qj2nNDNEBiysNi++B7YfSmgBUgSBQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:13:00.428061Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.09113","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:041b22c07e1a17a1e24cad5d8722d275f3bcdd100afcf6d33e6e96fb23ece29b","sha256:403db6a2404a171dbf8777589de74887e7536461d975481aef38f152fbbc6dda"],"state_sha256":"41f77b1f35c9acaaa553ce703eeceb9389b7720706a7448982813cf42572c249"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y79Vq3tBFDNu1KwPTQ1TfYrmfs9Px9QLU2RAtvVyI+TRM5Robo5iLsETsEG+rpO8LF5AakvovVB9EuxxgxxxCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T10:03:34.860356Z","bundle_sha256":"d900214c8d97a3737f5318c037ad2f7f8b42843b1f0764dfb4c512eda20fec62"}}