{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4EIWVIGMJ3WM36EBMYNUAHR4QM","short_pith_number":"pith:4EIWVIGM","schema_version":"1.0","canonical_sha256":"e1116aa0cc4eeccdf881661b401e3c8328522ee5331dd076385c1b69f42c8ae2","source":{"kind":"arxiv","id":"2502.13162","version":1},"attestation_state":"computed","paper":{"title":"ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Hao Wang, Huacan Wang, Ziyi Ni","submitted_at":"2025-02-16T18:47:41Z","abstract_excerpt":"Large Language Models (LLMs) have achieved remarkable success in various domains but remain vulnerable to adversarial jailbreak attacks. Existing prompt-defense strategies, including parameter-modifying and parameter-free approaches, face limitations in adaptability, interpretability, and customization, constraining their effectiveness against evolving threats. To address these challenges, we propose ShieldLearner, a novel paradigm that mimics human learning in defense. Through trial and error, it autonomously distills attack signatures into a Pattern Atlas and synthesizes defense heuristics i"},"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":"2502.13162","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CR","submitted_at":"2025-02-16T18:47:41Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"36989baaac9ff08ce2eb0de2a38ee3d7569b584618176cd94af85fd59dc54169","abstract_canon_sha256":"5092d030ff1eaba4c3bf017485d520b54dfd5a3fd98803ea21f466eb8d37f9f4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:16:43.273449Z","signature_b64":"fgHgHmpH+/7aJuGaiLBmsNBIigtxZIsPk3/IYyMOZcFdHpH8OJLZnXwROwLlzm4HtPKXIN9MuTCH1nidpXXLAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e1116aa0cc4eeccdf881661b401e3c8328522ee5331dd076385c1b69f42c8ae2","last_reissued_at":"2026-07-05T10:16:43.272962Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:16:43.272962Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ShieldLearner: A New Paradigm for Jailbreak Attack Defense in LLMs","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Hao Wang, Huacan Wang, Ziyi Ni","submitted_at":"2025-02-16T18:47:41Z","abstract_excerpt":"Large Language Models (LLMs) have achieved remarkable success in various domains but remain vulnerable to adversarial jailbreak attacks. Existing prompt-defense strategies, including parameter-modifying and parameter-free approaches, face limitations in adaptability, interpretability, and customization, constraining their effectiveness against evolving threats. To address these challenges, we propose ShieldLearner, a novel paradigm that mimics human learning in defense. Through trial and error, it autonomously distills attack signatures into a Pattern Atlas and synthesizes defense heuristics i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.13162","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/2502.13162/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":"2502.13162","created_at":"2026-07-05T10:16:43.273019+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.13162v1","created_at":"2026-07-05T10:16:43.273019+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.13162","created_at":"2026-07-05T10:16:43.273019+00:00"},{"alias_kind":"pith_short_12","alias_value":"4EIWVIGMJ3WM","created_at":"2026-07-05T10:16:43.273019+00:00"},{"alias_kind":"pith_short_16","alias_value":"4EIWVIGMJ3WM36EB","created_at":"2026-07-05T10:16:43.273019+00:00"},{"alias_kind":"pith_short_8","alias_value":"4EIWVIGM","created_at":"2026-07-05T10:16:43.273019+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2503.21460","citing_title":"Large Language Model Agent: A Survey on Methodology, Applications and Challenges","ref_index":190,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12548","citing_title":"DeepSeek Robustness Against Semantic-Character Dual-Space Mutated Prompt Injection","ref_index":19,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM","json":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM.json","graph_json":"https://pith.science/api/pith-number/4EIWVIGMJ3WM36EBMYNUAHR4QM/graph.json","events_json":"https://pith.science/api/pith-number/4EIWVIGMJ3WM36EBMYNUAHR4QM/events.json","paper":"https://pith.science/paper/4EIWVIGM"},"agent_actions":{"view_html":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM","download_json":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM.json","view_paper":"https://pith.science/paper/4EIWVIGM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.13162&json=true","fetch_graph":"https://pith.science/api/pith-number/4EIWVIGMJ3WM36EBMYNUAHR4QM/graph.json","fetch_events":"https://pith.science/api/pith-number/4EIWVIGMJ3WM36EBMYNUAHR4QM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM/action/storage_attestation","attest_author":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM/action/author_attestation","sign_citation":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM/action/citation_signature","submit_replication":"https://pith.science/pith/4EIWVIGMJ3WM36EBMYNUAHR4QM/action/replication_record"}},"created_at":"2026-07-05T10:16:43.273019+00:00","updated_at":"2026-07-05T10:16:43.273019+00:00"}