{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:37XZQINPDUH3GXFXRGSEDK6QFS","short_pith_number":"pith:37XZQINP","canonical_record":{"source":{"id":"2405.12604","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-21T08:57:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bb375c29a2eb445e1ed1b3a606c5a3b6591460a68458db9a3a2cedb3dfceefe7","abstract_canon_sha256":"2a9337aa1360bd92b02d1e6ae66c15ab666cd6ab1f4217a782cc4db4214736a6"},"schema_version":"1.0"},"canonical_sha256":"dfef9821af1d0fb35cb789a441abd02c92d53469fb55e2f5cea7474a235fc8fb","source":{"kind":"arxiv","id":"2405.12604","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.12604","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"arxiv_version","alias_value":"2405.12604v2","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.12604","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_12","alias_value":"37XZQINPDUH3","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_16","alias_value":"37XZQINPDUH3GXFX","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_8","alias_value":"37XZQINP","created_at":"2026-07-05T08:33:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:37XZQINPDUH3GXFXRGSEDK6QFS","target":"record","payload":{"canonical_record":{"source":{"id":"2405.12604","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-21T08:57:44Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"bb375c29a2eb445e1ed1b3a606c5a3b6591460a68458db9a3a2cedb3dfceefe7","abstract_canon_sha256":"2a9337aa1360bd92b02d1e6ae66c15ab666cd6ab1f4217a782cc4db4214736a6"},"schema_version":"1.0"},"canonical_sha256":"dfef9821af1d0fb35cb789a441abd02c92d53469fb55e2f5cea7474a235fc8fb","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:33:01.961509Z","signature_b64":"1WA5vunK90Kf4sACYxcv1+QDOXKcx68p+lKsGucN8BkzI6wBLGIY9oFoVwCOkX0fFDKqDXcjBdhBU9WQKtiGBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dfef9821af1d0fb35cb789a441abd02c92d53469fb55e2f5cea7474a235fc8fb","last_reissued_at":"2026-07-05T08:33:01.961009Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:33:01.961009Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2405.12604","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":"2026-07-05T08:33:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"962chB+XbOxmL826XgBG7yoki3jVNiW49i/ns+Ae1ojslpaJkH3gpmoi4b0E3QzHjuMZnXN/rWiRhFWF60lWBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T21:14:21.705987Z"},"content_sha256":"6d1d5c976478e882ca116e61d878661371f302c386d0d8969eca44fab0ae01cc","schema_version":"1.0","event_id":"sha256:6d1d5c976478e882ca116e61d878661371f302c386d0d8969eca44fab0ae01cc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:37XZQINPDUH3GXFXRGSEDK6QFS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Tiny Refinements Elicit Resilience: Toward Efficient Prefix-Model Against LLM Red-Teaming","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Jianhong Wang, Jiaxu Liu, Meng Fang, Sihao Wu, Xiangyu Yin, Xiaowei Huang, Xinping Yi","submitted_at":"2024-05-21T08:57:44Z","abstract_excerpt":"With the proliferation of red-teaming strategies for Large Language Models (LLMs), the deficiency in the literature about improving the safety and robustness of LLM defense strategies is becoming increasingly pronounced. This paper introduces the LLM-based \\textbf{sentinel} model as a plug-and-play prefix module designed to reconstruct the input prompt with just a few ($<30$) additional tokens, effectively reducing toxicity in responses from target LLMs. The sentinel model naturally overcomes the \\textit{parameter inefficiency} and \\textit{limited model accessibility} for fine-tuning large tar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.12604","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/2405.12604/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":"2026-07-05T08:33:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"E6UorNibW60SXY65N4+hW4QD7yDUEEBg9KiWYxA95UWZz6v5hw6k7HfMY0tzal8bTITow4SW1X8zQ5moiAZfBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-21T21:14:21.707003Z"},"content_sha256":"309aee5edf47abd8f85baf971114bbfe0d9b61d639a7e0a377f85c92f42524af","schema_version":"1.0","event_id":"sha256:309aee5edf47abd8f85baf971114bbfe0d9b61d639a7e0a377f85c92f42524af"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/37XZQINPDUH3GXFXRGSEDK6QFS/bundle.json","state_url":"https://pith.science/pith/37XZQINPDUH3GXFXRGSEDK6QFS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/37XZQINPDUH3GXFXRGSEDK6QFS/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":"2026-08-21T21:14:21Z","links":{"resolver":"https://pith.science/pith/37XZQINPDUH3GXFXRGSEDK6QFS","bundle":"https://pith.science/pith/37XZQINPDUH3GXFXRGSEDK6QFS/bundle.json","state":"https://pith.science/pith/37XZQINPDUH3GXFXRGSEDK6QFS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/37XZQINPDUH3GXFXRGSEDK6QFS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:37XZQINPDUH3GXFXRGSEDK6QFS","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":"2a9337aa1360bd92b02d1e6ae66c15ab666cd6ab1f4217a782cc4db4214736a6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-21T08:57:44Z","title_canon_sha256":"bb375c29a2eb445e1ed1b3a606c5a3b6591460a68458db9a3a2cedb3dfceefe7"},"schema_version":"1.0","source":{"id":"2405.12604","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2405.12604","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"arxiv_version","alias_value":"2405.12604v2","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.12604","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_12","alias_value":"37XZQINPDUH3","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_16","alias_value":"37XZQINPDUH3GXFX","created_at":"2026-07-05T08:33:01Z"},{"alias_kind":"pith_short_8","alias_value":"37XZQINP","created_at":"2026-07-05T08:33:01Z"}],"graph_snapshots":[{"event_id":"sha256:309aee5edf47abd8f85baf971114bbfe0d9b61d639a7e0a377f85c92f42524af","target":"graph","created_at":"2026-07-05T08:33:01Z","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/2405.12604/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"With the proliferation of red-teaming strategies for Large Language Models (LLMs), the deficiency in the literature about improving the safety and robustness of LLM defense strategies is becoming increasingly pronounced. This paper introduces the LLM-based \\textbf{sentinel} model as a plug-and-play prefix module designed to reconstruct the input prompt with just a few ($<30$) additional tokens, effectively reducing toxicity in responses from target LLMs. The sentinel model naturally overcomes the \\textit{parameter inefficiency} and \\textit{limited model accessibility} for fine-tuning large tar","authors_text":"Jianhong Wang, Jiaxu Liu, Meng Fang, Sihao Wu, Xiangyu Yin, Xiaowei Huang, Xinping Yi","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-21T08:57:44Z","title":"Tiny Refinements Elicit Resilience: Toward Efficient Prefix-Model Against LLM Red-Teaming"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.12604","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":"sha256:6d1d5c976478e882ca116e61d878661371f302c386d0d8969eca44fab0ae01cc","target":"record","created_at":"2026-07-05T08:33:01Z","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":"2a9337aa1360bd92b02d1e6ae66c15ab666cd6ab1f4217a782cc4db4214736a6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-05-21T08:57:44Z","title_canon_sha256":"bb375c29a2eb445e1ed1b3a606c5a3b6591460a68458db9a3a2cedb3dfceefe7"},"schema_version":"1.0","source":{"id":"2405.12604","kind":"arxiv","version":2}},"canonical_sha256":"dfef9821af1d0fb35cb789a441abd02c92d53469fb55e2f5cea7474a235fc8fb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dfef9821af1d0fb35cb789a441abd02c92d53469fb55e2f5cea7474a235fc8fb","first_computed_at":"2026-07-05T08:33:01.961009Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:33:01.961009Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"1WA5vunK90Kf4sACYxcv1+QDOXKcx68p+lKsGucN8BkzI6wBLGIY9oFoVwCOkX0fFDKqDXcjBdhBU9WQKtiGBw==","signature_status":"signed_v1","signed_at":"2026-07-05T08:33:01.961509Z","signed_message":"canonical_sha256_bytes"},"source_id":"2405.12604","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6d1d5c976478e882ca116e61d878661371f302c386d0d8969eca44fab0ae01cc","sha256:309aee5edf47abd8f85baf971114bbfe0d9b61d639a7e0a377f85c92f42524af"],"state_sha256":"e388af31cefc76c29bd827d2d67defeec7af1bc59ce24dd1561be2c192feaf41"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"wT8PLFDMdABNFSw1o97UEfwJD4jnUw687zWg0lA2WCJmcJy+/+SysZNU8lW2pfLb7EfrVz7PBwNhwrvzkJhdCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-21T21:14:21.715200Z","bundle_sha256":"06570cb387d78fd9684cd14448f692794544e54052f2659842de481ac5aa15aa"}}