{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DX7LCQUNGS5FAEN56TPU4OCMAS","short_pith_number":"pith:DX7LCQUN","schema_version":"1.0","canonical_sha256":"1dfeb1428d34ba5011bdf4df4e384c04a4ebdb347ab3a4e040b322f712b18c5b","source":{"kind":"arxiv","id":"2406.06852","version":5},"attestation_state":"computed","paper":{"title":"A Survey of Recent Backdoor Attacks and Defenses in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Fengjun Pan, Jie Fu, Leilei Gan, Luu Anh Tuan, Meihuizi Jia, Shuai Zhao, Xiaobao Wu, Xiaoyu Xu, Yichao Feng, Zhongliang Guo","submitted_at":"2024-06-10T23:54:21Z","abstract_excerpt":"Large Language Models (LLMs), which bridge the gap between human language understanding and complex problem-solving, achieve state-of-the-art performance on several NLP tasks, particularly in few-shot and zero-shot settings. Despite the demonstrable efficacy of LLMs, due to constraints on computational resources, users have to engage with open-source language models or outsource the entire training process to third-party platforms. However, research has demonstrated that language models are susceptible to potential security vulnerabilities, particularly in backdoor attacks. Backdoor attacks ar"},"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":"2406.06852","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2024-06-10T23:54:21Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"6227480b4735cf03345da47305afa3d45822c8661bfb8f6c64297ec52d160326","abstract_canon_sha256":"e25a51ff50f1c843f2b552d89f095c8e95e114c9a79485fd80ae29edc78af73c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:56:52.420398Z","signature_b64":"GcZkyOwr7l77/k0CyPberTrFy7o59ma0cbnipHuK/q2RFMnhOwbSTZBjynB61hg6UFN9P12CR41UuiZLWYceBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1dfeb1428d34ba5011bdf4df4e384c04a4ebdb347ab3a4e040b322f712b18c5b","last_reissued_at":"2026-07-05T09:56:52.419887Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:56:52.419887Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Survey of Recent Backdoor Attacks and Defenses in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Fengjun Pan, Jie Fu, Leilei Gan, Luu Anh Tuan, Meihuizi Jia, Shuai Zhao, Xiaobao Wu, Xiaoyu Xu, Yichao Feng, Zhongliang Guo","submitted_at":"2024-06-10T23:54:21Z","abstract_excerpt":"Large Language Models (LLMs), which bridge the gap between human language understanding and complex problem-solving, achieve state-of-the-art performance on several NLP tasks, particularly in few-shot and zero-shot settings. Despite the demonstrable efficacy of LLMs, due to constraints on computational resources, users have to engage with open-source language models or outsource the entire training process to third-party platforms. However, research has demonstrated that language models are susceptible to potential security vulnerabilities, particularly in backdoor attacks. Backdoor attacks ar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.06852","kind":"arxiv","version":5},"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/2406.06852/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":"2406.06852","created_at":"2026-07-05T09:56:52.419947+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.06852v5","created_at":"2026-07-05T09:56:52.419947+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.06852","created_at":"2026-07-05T09:56:52.419947+00:00"},{"alias_kind":"pith_short_12","alias_value":"DX7LCQUNGS5F","created_at":"2026-07-05T09:56:52.419947+00:00"},{"alias_kind":"pith_short_16","alias_value":"DX7LCQUNGS5FAEN5","created_at":"2026-07-05T09:56:52.419947+00:00"},{"alias_kind":"pith_short_8","alias_value":"DX7LCQUN","created_at":"2026-07-05T09:56:52.419947+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.31309","citing_title":"CSO-LLM: Class Subspace Orthogonalization for Post-Training Backdoor Detection and Trigger Inversion in LLMs","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2409.18169","citing_title":"Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey","ref_index":180,"is_internal_anchor":false},{"citing_arxiv_id":"2501.18416","citing_title":"Exploring Potential Prompt Injection Attacks in Federated Military LLMs and Their Mitigation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2502.20650","citing_title":"Gungnir: Exploiting Stylistic Features in Images for Backdoor Attacks on Diffusion Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2510.18333","citing_title":"Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02255","citing_title":"On the Privacy of LLMs: An Ablation Study","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS","json":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS.json","graph_json":"https://pith.science/api/pith-number/DX7LCQUNGS5FAEN56TPU4OCMAS/graph.json","events_json":"https://pith.science/api/pith-number/DX7LCQUNGS5FAEN56TPU4OCMAS/events.json","paper":"https://pith.science/paper/DX7LCQUN"},"agent_actions":{"view_html":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS","download_json":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS.json","view_paper":"https://pith.science/paper/DX7LCQUN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.06852&json=true","fetch_graph":"https://pith.science/api/pith-number/DX7LCQUNGS5FAEN56TPU4OCMAS/graph.json","fetch_events":"https://pith.science/api/pith-number/DX7LCQUNGS5FAEN56TPU4OCMAS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS/action/storage_attestation","attest_author":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS/action/author_attestation","sign_citation":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS/action/citation_signature","submit_replication":"https://pith.science/pith/DX7LCQUNGS5FAEN56TPU4OCMAS/action/replication_record"}},"created_at":"2026-07-05T09:56:52.419947+00:00","updated_at":"2026-07-05T09:56:52.419947+00:00"}