{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QNIDVU2M4PBEXMDYBI4FTDQRAW","short_pith_number":"pith:QNIDVU2M","schema_version":"1.0","canonical_sha256":"83503ad34ce3c24bb0780a38598e1105a37e6c20e2ec3748b56d86b5f28f9f49","source":{"kind":"arxiv","id":"2402.12168","version":3},"attestation_state":"computed","paper":{"title":"Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Jie Fu, Jinming Wen, Leilei Gan, Lingjuan Lyu, Luu Anh Tuan, Meihuizi Jia, Shuai Zhao","submitted_at":"2024-02-19T14:22:54Z","abstract_excerpt":"Recently, various parameter-efficient fine-tuning (PEFT) strategies for application to language models have been proposed and successfully implemented. However, this raises the question of whether PEFT, which only updates a limited set of model parameters, constitutes security vulnerabilities when confronted with weight-poisoning backdoor attacks. In this study, we show that PEFT is more susceptible to weight-poisoning backdoor attacks compared to the full-parameter fine-tuning method, with pre-defined triggers remaining exploitable and pre-defined targets maintaining high confidence, even aft"},"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":"2402.12168","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-02-19T14:22:54Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"1f89f6d4db00a1f00fab264594562aa267bd3d8de1ad6e78a2f07c578a952c27","abstract_canon_sha256":"de0ea7b5b6d75f970f92be86d03e04a3623e39dec07b63fca52a33072a16b69e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:02:05.045703Z","signature_b64":"6zbrdFHHldV5gWEnmDbGrRWT2DgWo0PI+snYKF/ltxLWsG9Xf8eDUofA5jtCy2+AcaXsfwQ1hrmw8J2D4EvMAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"83503ad34ce3c24bb0780a38598e1105a37e6c20e2ec3748b56d86b5f28f9f49","last_reissued_at":"2026-07-05T08:02:05.045207Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:02:05.045207Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Defending Against Weight-Poisoning Backdoor Attacks for Parameter-Efficient Fine-Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Jie Fu, Jinming Wen, Leilei Gan, Lingjuan Lyu, Luu Anh Tuan, Meihuizi Jia, Shuai Zhao","submitted_at":"2024-02-19T14:22:54Z","abstract_excerpt":"Recently, various parameter-efficient fine-tuning (PEFT) strategies for application to language models have been proposed and successfully implemented. However, this raises the question of whether PEFT, which only updates a limited set of model parameters, constitutes security vulnerabilities when confronted with weight-poisoning backdoor attacks. In this study, we show that PEFT is more susceptible to weight-poisoning backdoor attacks compared to the full-parameter fine-tuning method, with pre-defined triggers remaining exploitable and pre-defined targets maintaining high confidence, even aft"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.12168","kind":"arxiv","version":3},"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/2402.12168/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":"2402.12168","created_at":"2026-07-05T08:02:05.045271+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.12168v3","created_at":"2026-07-05T08:02:05.045271+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.12168","created_at":"2026-07-05T08:02:05.045271+00:00"},{"alias_kind":"pith_short_12","alias_value":"QNIDVU2M4PBE","created_at":"2026-07-05T08:02:05.045271+00:00"},{"alias_kind":"pith_short_16","alias_value":"QNIDVU2M4PBEXMDY","created_at":"2026-07-05T08:02:05.045271+00:00"},{"alias_kind":"pith_short_8","alias_value":"QNIDVU2M","created_at":"2026-07-05T08:02:05.045271+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.24350","citing_title":"Unveiling the Backdoor Mechanism Hidden Behind Catastrophic Overfitting in Fast Adversarial Training","ref_index":97,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04738","citing_title":"Fine-Tuning Integrity for Modern Neural Networks: Structured Drift Proofs via Norm, Rank, and Sparsity Certificates","ref_index":45,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW","json":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW.json","graph_json":"https://pith.science/api/pith-number/QNIDVU2M4PBEXMDYBI4FTDQRAW/graph.json","events_json":"https://pith.science/api/pith-number/QNIDVU2M4PBEXMDYBI4FTDQRAW/events.json","paper":"https://pith.science/paper/QNIDVU2M"},"agent_actions":{"view_html":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW","download_json":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW.json","view_paper":"https://pith.science/paper/QNIDVU2M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.12168&json=true","fetch_graph":"https://pith.science/api/pith-number/QNIDVU2M4PBEXMDYBI4FTDQRAW/graph.json","fetch_events":"https://pith.science/api/pith-number/QNIDVU2M4PBEXMDYBI4FTDQRAW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW/action/storage_attestation","attest_author":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW/action/author_attestation","sign_citation":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW/action/citation_signature","submit_replication":"https://pith.science/pith/QNIDVU2M4PBEXMDYBI4FTDQRAW/action/replication_record"}},"created_at":"2026-07-05T08:02:05.045271+00:00","updated_at":"2026-07-05T08:02:05.045271+00:00"}