{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:STC2A4WWIJKS7G6IZCI75AMFJX","short_pith_number":"pith:STC2A4WW","schema_version":"1.0","canonical_sha256":"94c5a072d642552f9bc8c891fe81854dc38d6f0598d0dab6d0bee4b4d5534016","source":{"kind":"arxiv","id":"2402.15727","version":2},"attestation_state":"computed","paper":{"title":"LLMs Can Defend Themselves Against Jailbreaking in a Practical Manner: A Vision Paper","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Daoyuan Wu, Ning Liu, Shuai Wang, Yang Liu","submitted_at":"2024-02-24T05:34:43Z","abstract_excerpt":"Jailbreaking is an emerging adversarial attack that bypasses the safety alignment deployed in off-the-shelf large language models (LLMs). A considerable amount of research exists proposing more effective jailbreak attacks, including the recent Greedy Coordinate Gradient (GCG) attack, jailbreak template-based attacks such as using \"Do-Anything-Now\" (DAN), and multilingual jailbreak. In contrast, the defensive side has been relatively less explored. This paper proposes a lightweight yet practical defense called SELFDEFEND, which can defend against all existing jailbreak attacks with minimal dela"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2402.15727","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-02-24T05:34:43Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"492fdfafbaeed0b5b06a47fe9daf92a1cab7b3a0b74f57ecf8edb5cc18189b45","abstract_canon_sha256":"695eed76849b1dd3ddbf713e2eee8a0bf5d87d173d5b5d9ceb8e98e931c2f98b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:51:39.858028Z","signature_b64":"XZ98TWw3DbSH+Lb80iSi1yicrMILUvrtu/505OQR8WQOQxcVDfPiJcQ3fW9gqJNFORzu0rW4iQc3sY1lCu1uCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"94c5a072d642552f9bc8c891fe81854dc38d6f0598d0dab6d0bee4b4d5534016","last_reissued_at":"2026-07-05T07:51:39.857598Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:51:39.857598Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLMs Can Defend Themselves Against Jailbreaking in a Practical Manner: A Vision Paper","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Daoyuan Wu, Ning Liu, Shuai Wang, Yang Liu","submitted_at":"2024-02-24T05:34:43Z","abstract_excerpt":"Jailbreaking is an emerging adversarial attack that bypasses the safety alignment deployed in off-the-shelf large language models (LLMs). A considerable amount of research exists proposing more effective jailbreak attacks, including the recent Greedy Coordinate Gradient (GCG) attack, jailbreak template-based attacks such as using \"Do-Anything-Now\" (DAN), and multilingual jailbreak. In contrast, the defensive side has been relatively less explored. This paper proposes a lightweight yet practical defense called SELFDEFEND, which can defend against all existing jailbreak attacks with minimal dela"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.15727","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/2402.15727/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2402.15727","created_at":"2026-07-05T07:51:39.857660+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.15727v2","created_at":"2026-07-05T07:51:39.857660+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.15727","created_at":"2026-07-05T07:51:39.857660+00:00"},{"alias_kind":"pith_short_12","alias_value":"STC2A4WWIJKS","created_at":"2026-07-05T07:51:39.857660+00:00"},{"alias_kind":"pith_short_16","alias_value":"STC2A4WWIJKS7G6I","created_at":"2026-07-05T07:51:39.857660+00:00"},{"alias_kind":"pith_short_8","alias_value":"STC2A4WW","created_at":"2026-07-05T07:51:39.857660+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2506.01770","citing_title":"ReGA: Model-Based Safeguard for LLMs via Representation-Guided Abstraction","ref_index":89,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX","json":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX.json","graph_json":"https://pith.science/api/pith-number/STC2A4WWIJKS7G6IZCI75AMFJX/graph.json","events_json":"https://pith.science/api/pith-number/STC2A4WWIJKS7G6IZCI75AMFJX/events.json","paper":"https://pith.science/paper/STC2A4WW"},"agent_actions":{"view_html":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX","download_json":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX.json","view_paper":"https://pith.science/paper/STC2A4WW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.15727&json=true","fetch_graph":"https://pith.science/api/pith-number/STC2A4WWIJKS7G6IZCI75AMFJX/graph.json","fetch_events":"https://pith.science/api/pith-number/STC2A4WWIJKS7G6IZCI75AMFJX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX/action/storage_attestation","attest_author":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX/action/author_attestation","sign_citation":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX/action/citation_signature","submit_replication":"https://pith.science/pith/STC2A4WWIJKS7G6IZCI75AMFJX/action/replication_record"}},"created_at":"2026-07-05T07:51:39.857660+00:00","updated_at":"2026-07-05T07:51:39.857660+00:00"}