{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:J6AGAQKKQU5YZBBWZGUARDVIXA","short_pith_number":"pith:J6AGAQKK","canonical_record":{"source":{"id":"2506.23576","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-30T07:29:07Z","cross_cats_sorted":[],"title_canon_sha256":"962395f2cdcf88e96dc5d115522eae8a7916fcebedb9d3ec152322e69e574032","abstract_canon_sha256":"6fd507fa2e31329afeaf6b3536cddcfc4bd94e882040a23d0144464f88dfcc48"},"schema_version":"1.0"},"canonical_sha256":"4f8060414a853b8c8436c9a8088ea8b823ed9a1d6913da3bd11e524e26e13caf","source":{"kind":"arxiv","id":"2506.23576","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.23576","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"arxiv_version","alias_value":"2506.23576v1","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.23576","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_12","alias_value":"J6AGAQKKQU5Y","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_16","alias_value":"J6AGAQKKQU5YZBBW","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_8","alias_value":"J6AGAQKK","created_at":"2026-07-05T11:29:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:J6AGAQKKQU5YZBBWZGUARDVIXA","target":"record","payload":{"canonical_record":{"source":{"id":"2506.23576","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-30T07:29:07Z","cross_cats_sorted":[],"title_canon_sha256":"962395f2cdcf88e96dc5d115522eae8a7916fcebedb9d3ec152322e69e574032","abstract_canon_sha256":"6fd507fa2e31329afeaf6b3536cddcfc4bd94e882040a23d0144464f88dfcc48"},"schema_version":"1.0"},"canonical_sha256":"4f8060414a853b8c8436c9a8088ea8b823ed9a1d6913da3bd11e524e26e13caf","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:29:24.276182Z","signature_b64":"DTUfNsL35QOcmgrG3bEIqbDfHW3EkmrFASoB32ZnrZas4bANYsSdMRDtdOGb6yLyU8VXBY+NIa1veqlqr6U7CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4f8060414a853b8c8436c9a8088ea8b823ed9a1d6913da3bd11e524e26e13caf","last_reissued_at":"2026-07-05T11:29:24.275632Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:29:24.275632Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.23576","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:29:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pfAQ1W7EwOReg1tArqM4g1Pci3r95tJYh4l/2JX2wrer3sf1+vlb8GMbk5lpgH6+EzJdQ4Ns4eZt9cwdGQJFDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T17:50:35.789356Z"},"content_sha256":"e8db4ad218a3296b5aeb40c3b61e554ff95038926c78957166f8c18970754b0d","schema_version":"1.0","event_id":"sha256:e8db4ad218a3296b5aeb40c3b61e554ff95038926c78957166f8c18970754b0d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:J6AGAQKKQU5YZBBWZGUARDVIXA","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Evaluating Multi-Agent Defences Against Jailbreaking Attacks on Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jun Pang, Maria Carolina Cornelia Wit","submitted_at":"2025-06-30T07:29:07Z","abstract_excerpt":"Recent advances in large language models (LLMs) have raised concerns about jailbreaking attacks, i.e., prompts that bypass safety mechanisms. This paper investigates the use of multi-agent LLM systems as a defence against such attacks. We evaluate three jailbreaking strategies, including the original AutoDefense attack and two from Deepleaps: BetterDan and JB. Reproducing the AutoDefense framework, we compare single-agent setups with two- and three-agent configurations. Our results show that multi-agent systems enhance resistance to jailbreaks, especially by reducing false negatives. However, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.23576","kind":"arxiv","version":1},"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/2506.23576/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":"2026-07-05T11:29:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"TZqs5wDPb+hhuolAJRhj4gcnHUjV1teWhsWJpbo4lShH8kWf8L+rDpmbpuoYvwO7T9c7YFNMSKQOKk2KzW4hDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-16T17:50:35.790333Z"},"content_sha256":"3e295fd764622654b8d695e40dd9684a49187dadb0913dbd445c7ffdd5f07057","schema_version":"1.0","event_id":"sha256:3e295fd764622654b8d695e40dd9684a49187dadb0913dbd445c7ffdd5f07057"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/bundle.json","state_url":"https://pith.science/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/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":"2026-08-16T17:50:35Z","links":{"resolver":"https://pith.science/pith/J6AGAQKKQU5YZBBWZGUARDVIXA","bundle":"https://pith.science/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/bundle.json","state":"https://pith.science/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/state.json","well_known_bundle":"https://pith.science/.well-known/pith/J6AGAQKKQU5YZBBWZGUARDVIXA/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:J6AGAQKKQU5YZBBWZGUARDVIXA","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":"6fd507fa2e31329afeaf6b3536cddcfc4bd94e882040a23d0144464f88dfcc48","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-30T07:29:07Z","title_canon_sha256":"962395f2cdcf88e96dc5d115522eae8a7916fcebedb9d3ec152322e69e574032"},"schema_version":"1.0","source":{"id":"2506.23576","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.23576","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"arxiv_version","alias_value":"2506.23576v1","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.23576","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_12","alias_value":"J6AGAQKKQU5Y","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_16","alias_value":"J6AGAQKKQU5YZBBW","created_at":"2026-07-05T11:29:24Z"},{"alias_kind":"pith_short_8","alias_value":"J6AGAQKK","created_at":"2026-07-05T11:29:24Z"}],"graph_snapshots":[{"event_id":"sha256:3e295fd764622654b8d695e40dd9684a49187dadb0913dbd445c7ffdd5f07057","target":"graph","created_at":"2026-07-05T11:29:24Z","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/2506.23576/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances in large language models (LLMs) have raised concerns about jailbreaking attacks, i.e., prompts that bypass safety mechanisms. This paper investigates the use of multi-agent LLM systems as a defence against such attacks. We evaluate three jailbreaking strategies, including the original AutoDefense attack and two from Deepleaps: BetterDan and JB. Reproducing the AutoDefense framework, we compare single-agent setups with two- and three-agent configurations. Our results show that multi-agent systems enhance resistance to jailbreaks, especially by reducing false negatives. However, ","authors_text":"Jun Pang, Maria Carolina Cornelia Wit","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-30T07:29:07Z","title":"Evaluating Multi-Agent Defences Against Jailbreaking Attacks on Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.23576","kind":"arxiv","version":1},"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:e8db4ad218a3296b5aeb40c3b61e554ff95038926c78957166f8c18970754b0d","target":"record","created_at":"2026-07-05T11:29:24Z","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":"6fd507fa2e31329afeaf6b3536cddcfc4bd94e882040a23d0144464f88dfcc48","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-30T07:29:07Z","title_canon_sha256":"962395f2cdcf88e96dc5d115522eae8a7916fcebedb9d3ec152322e69e574032"},"schema_version":"1.0","source":{"id":"2506.23576","kind":"arxiv","version":1}},"canonical_sha256":"4f8060414a853b8c8436c9a8088ea8b823ed9a1d6913da3bd11e524e26e13caf","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4f8060414a853b8c8436c9a8088ea8b823ed9a1d6913da3bd11e524e26e13caf","first_computed_at":"2026-07-05T11:29:24.275632Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:29:24.275632Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"DTUfNsL35QOcmgrG3bEIqbDfHW3EkmrFASoB32ZnrZas4bANYsSdMRDtdOGb6yLyU8VXBY+NIa1veqlqr6U7CQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:29:24.276182Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.23576","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e8db4ad218a3296b5aeb40c3b61e554ff95038926c78957166f8c18970754b0d","sha256:3e295fd764622654b8d695e40dd9684a49187dadb0913dbd445c7ffdd5f07057"],"state_sha256":"0f2461ca31b53be88712ca4671b25a3475a03ef7ad90cf2bc6d55c7d257a60c1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8qCiyhEYtp+mMdmHqjNgGCtzCrNL6kKXVpSNTo3GX+mtyV/UddI7oUq8Krmi23WWvZGKBdZ+//ggVXEe0FGuAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-16T17:50:35.864198Z","bundle_sha256":"58533eae4e0e6899503264fb80526f6610e62ea611fd70b9306f8dc837ba2c2b"}}