{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:Z3YHUWECHAILUPMF5KFFWWIHZI","short_pith_number":"pith:Z3YHUWEC","canonical_record":{"source":{"id":"2405.20770","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T07:23:56Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"74064baa5b25eda053d343eb07b5ba39d263ce4f28e7da82a150e75677a99427","abstract_canon_sha256":"e6b7961b81f30d9736f6fdfe68afef0fe6d869a59a25c8d0a603b072f4d1bb14"},"schema_version":"1.0"},"canonical_sha256":"cef07a58823810ba3d85ea8a5b5907ca2dc1557b77f74b8aaca2111ceb9d5b1a","source":{"kind":"arxiv","id":"2405.20770","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.20770","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"arxiv_version","alias_value":"2405.20770v4","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20770","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_12","alias_value":"Z3YHUWECHAIL","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_16","alias_value":"Z3YHUWECHAILUPMF","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_8","alias_value":"Z3YHUWEC","created_at":"2026-07-05T10:52:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:Z3YHUWECHAILUPMF5KFFWWIHZI","target":"record","payload":{"canonical_record":{"source":{"id":"2405.20770","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T07:23:56Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"74064baa5b25eda053d343eb07b5ba39d263ce4f28e7da82a150e75677a99427","abstract_canon_sha256":"e6b7961b81f30d9736f6fdfe68afef0fe6d869a59a25c8d0a603b072f4d1bb14"},"schema_version":"1.0"},"canonical_sha256":"cef07a58823810ba3d85ea8a5b5907ca2dc1557b77f74b8aaca2111ceb9d5b1a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:52:36.960134Z","signature_b64":"vqy+fDA/VfH/LepAQPmCpy1CN0xRxdtpyzipHvuNEEILSZBKRzJ1p+OG4BoxdNkuxHU5EkNr1BjiQBQrFIktCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cef07a58823810ba3d85ea8a5b5907ca2dc1557b77f74b8aaca2111ceb9d5b1a","last_reissued_at":"2026-07-05T10:52:36.959659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:52:36.959659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.20770","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":"2026-07-05T10:52:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QAjrDaCLrAYNmFKA7sMQMYRdqXlqj3m+6xtL38Rc513TcWUoH3ggZd8+UdMPxiMYFiAaDQiB43/1F2hHHn6pBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T17:33:48.466216Z"},"content_sha256":"c6b73a655a8cbec789340e7b09fedd337e2c0f87303853201f8b7a4c5e68b02a","schema_version":"1.0","event_id":"sha256:c6b73a655a8cbec789340e7b09fedd337e2c0f87303853201f8b7a4c5e68b02a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:Z3YHUWECHAILUPMF5KFFWWIHZI","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Large Language Model Sentinel: LLM Agent for Adversarial Purification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.CL","authors_text":"Guang Lin, Qibin Zhao, Toshihisa Tanaka","submitted_at":"2024-05-24T07:23:56Z","abstract_excerpt":"Over the past two years, the use of large language models (LLMs) has advanced rapidly. While these LLMs offer considerable convenience, they also raise security concerns, as LLMs are vulnerable to adversarial attacks by some well-designed textual perturbations. In this paper, we introduce a novel defense technique named Large LAnguage MOdel Sentinel (LLAMOS), which is designed to enhance the adversarial robustness of LLMs by purifying the adversarial textual examples before feeding them into the target LLM. Our method comprises two main components: a) Agent instruction, which can simulate a ne"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20770","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/2405.20770/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":"2026-07-05T10:52:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Heyg0nr6a6UWVcOtlWmvjsGVAz6pTXcXdBjhBiSucYeTpnqBOCtdxHtodoiTiDbREC4p4DpYyEdmNIw8PQyKCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-15T17:33:48.466731Z"},"content_sha256":"d0409d10b24459e0120f7cd13d874ad8290b0065dd2d20540b9c5743f9480524","schema_version":"1.0","event_id":"sha256:d0409d10b24459e0120f7cd13d874ad8290b0065dd2d20540b9c5743f9480524"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/bundle.json","state_url":"https://pith.science/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/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":"2026-08-15T17:33:48Z","links":{"resolver":"https://pith.science/pith/Z3YHUWECHAILUPMF5KFFWWIHZI","bundle":"https://pith.science/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/bundle.json","state":"https://pith.science/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/state.json","well_known_bundle":"https://pith.science/.well-known/pith/Z3YHUWECHAILUPMF5KFFWWIHZI/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:Z3YHUWECHAILUPMF5KFFWWIHZI","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":"e6b7961b81f30d9736f6fdfe68afef0fe6d869a59a25c8d0a603b072f4d1bb14","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T07:23:56Z","title_canon_sha256":"74064baa5b25eda053d343eb07b5ba39d263ce4f28e7da82a150e75677a99427"},"schema_version":"1.0","source":{"id":"2405.20770","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.20770","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"arxiv_version","alias_value":"2405.20770v4","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.20770","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_12","alias_value":"Z3YHUWECHAIL","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_16","alias_value":"Z3YHUWECHAILUPMF","created_at":"2026-07-05T10:52:36Z"},{"alias_kind":"pith_short_8","alias_value":"Z3YHUWEC","created_at":"2026-07-05T10:52:36Z"}],"graph_snapshots":[{"event_id":"sha256:d0409d10b24459e0120f7cd13d874ad8290b0065dd2d20540b9c5743f9480524","target":"graph","created_at":"2026-07-05T10:52:36Z","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/2405.20770/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Over the past two years, the use of large language models (LLMs) has advanced rapidly. While these LLMs offer considerable convenience, they also raise security concerns, as LLMs are vulnerable to adversarial attacks by some well-designed textual perturbations. In this paper, we introduce a novel defense technique named Large LAnguage MOdel Sentinel (LLAMOS), which is designed to enhance the adversarial robustness of LLMs by purifying the adversarial textual examples before feeding them into the target LLM. Our method comprises two main components: a) Agent instruction, which can simulate a ne","authors_text":"Guang Lin, Qibin Zhao, Toshihisa Tanaka","cross_cats":["cs.AI","cs.CR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T07:23:56Z","title":"Large Language Model Sentinel: LLM Agent for Adversarial Purification"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.20770","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:c6b73a655a8cbec789340e7b09fedd337e2c0f87303853201f8b7a4c5e68b02a","target":"record","created_at":"2026-07-05T10:52:36Z","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":"e6b7961b81f30d9736f6fdfe68afef0fe6d869a59a25c8d0a603b072f4d1bb14","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-24T07:23:56Z","title_canon_sha256":"74064baa5b25eda053d343eb07b5ba39d263ce4f28e7da82a150e75677a99427"},"schema_version":"1.0","source":{"id":"2405.20770","kind":"arxiv","version":4}},"canonical_sha256":"cef07a58823810ba3d85ea8a5b5907ca2dc1557b77f74b8aaca2111ceb9d5b1a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cef07a58823810ba3d85ea8a5b5907ca2dc1557b77f74b8aaca2111ceb9d5b1a","first_computed_at":"2026-07-05T10:52:36.959659Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:52:36.959659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"vqy+fDA/VfH/LepAQPmCpy1CN0xRxdtpyzipHvuNEEILSZBKRzJ1p+OG4BoxdNkuxHU5EkNr1BjiQBQrFIktCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T10:52:36.960134Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.20770","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c6b73a655a8cbec789340e7b09fedd337e2c0f87303853201f8b7a4c5e68b02a","sha256:d0409d10b24459e0120f7cd13d874ad8290b0065dd2d20540b9c5743f9480524"],"state_sha256":"d508036fe4a786c01550a115219e8c63185dd4219957a537810ec14febf29367"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ThAhxh745bMzs2sW0LjMniR3PJ08mklna7BNVLs1vgm5BxMj/DgKZrVcusFeMFkS0Pmi8C1btS06Jv5pTVQfAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-15T17:33:48.471505Z","bundle_sha256":"8fa1558c058dce98edee521bae702252b2474e10d9fe1b496c9e2e5151a4785c"}}