{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SRPTMO2C5FU4H42BL4XJRMUTEQ","short_pith_number":"pith:SRPTMO2C","schema_version":"1.0","canonical_sha256":"945f363b42e969c3f3415f2e98b2932403856e416efc01119bd0c18632848748","source":{"kind":"arxiv","id":"2405.16833","version":2},"attestation_state":"computed","paper":{"title":"Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chia-Mu Yu, Chia-Yi Hsu, Chih-Hsun Lin, Chun-Ying Huang, Pin-Yu Chen, Yu-Lin Tsai","submitted_at":"2024-05-27T05:04:05Z","abstract_excerpt":"While large language models (LLMs) such as Llama-2 or GPT-4 have shown impressive zero-shot performance, fine-tuning is still necessary to enhance their performance for customized datasets, domain-specific tasks, or other private needs. However, fine-tuning all parameters of LLMs requires significant hardware resources, which can be impractical for typical users. Therefore, parameter-efficient fine-tuning such as LoRA have emerged, allowing users to fine-tune LLMs without the need for considerable computing resources, with little performance degradation compared to fine-tuning all parameters. "},"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":"2405.16833","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-05-27T05:04:05Z","cross_cats_sorted":[],"title_canon_sha256":"293fa5b6902b0b1fb3983d4f1cff2483142a2d07e1f9be00ed9218e5d68561fc","abstract_canon_sha256":"840943178a8e606db289fb2e46b86ea8f9e31c5d8041ddb7417c94a7802ee7b9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:56:51.809889Z","signature_b64":"TTRXandTFPG0HiznjaeZYdrNgHno5hH3Oe3Uqk+peJEDcju0h/krJ6W30WCb1ZpyHpCNVCCWFIhLhXVwK/tiDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"945f363b42e969c3f3415f2e98b2932403856e416efc01119bd0c18632848748","last_reissued_at":"2026-07-05T09:56:51.809404Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:56:51.809404Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Safe LoRA: the Silver Lining of Reducing Safety Risks when Fine-tuning Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chia-Mu Yu, Chia-Yi Hsu, Chih-Hsun Lin, Chun-Ying Huang, Pin-Yu Chen, Yu-Lin Tsai","submitted_at":"2024-05-27T05:04:05Z","abstract_excerpt":"While large language models (LLMs) such as Llama-2 or GPT-4 have shown impressive zero-shot performance, fine-tuning is still necessary to enhance their performance for customized datasets, domain-specific tasks, or other private needs. However, fine-tuning all parameters of LLMs requires significant hardware resources, which can be impractical for typical users. Therefore, parameter-efficient fine-tuning such as LoRA have emerged, allowing users to fine-tune LLMs without the need for considerable computing resources, with little performance degradation compared to fine-tuning all parameters. "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.16833","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/2405.16833/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":"2405.16833","created_at":"2026-07-05T09:56:51.809455+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.16833v2","created_at":"2026-07-05T09:56:51.809455+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.16833","created_at":"2026-07-05T09:56:51.809455+00:00"},{"alias_kind":"pith_short_12","alias_value":"SRPTMO2C5FU4","created_at":"2026-07-05T09:56:51.809455+00:00"},{"alias_kind":"pith_short_16","alias_value":"SRPTMO2C5FU4H42B","created_at":"2026-07-05T09:56:51.809455+00:00"},{"alias_kind":"pith_short_8","alias_value":"SRPTMO2C","created_at":"2026-07-05T09:56:51.809455+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2409.18169","citing_title":"Harmful Fine-tuning Attacks and Defenses for Large Language Models: A Survey","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17691","citing_title":"SafeAnchor: Preventing Cumulative Safety Erosion in Continual Domain Adaptation of Large Language Models","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ","json":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ.json","graph_json":"https://pith.science/api/pith-number/SRPTMO2C5FU4H42BL4XJRMUTEQ/graph.json","events_json":"https://pith.science/api/pith-number/SRPTMO2C5FU4H42BL4XJRMUTEQ/events.json","paper":"https://pith.science/paper/SRPTMO2C"},"agent_actions":{"view_html":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ","download_json":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ.json","view_paper":"https://pith.science/paper/SRPTMO2C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.16833&json=true","fetch_graph":"https://pith.science/api/pith-number/SRPTMO2C5FU4H42BL4XJRMUTEQ/graph.json","fetch_events":"https://pith.science/api/pith-number/SRPTMO2C5FU4H42BL4XJRMUTEQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ/action/storage_attestation","attest_author":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ/action/author_attestation","sign_citation":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ/action/citation_signature","submit_replication":"https://pith.science/pith/SRPTMO2C5FU4H42BL4XJRMUTEQ/action/replication_record"}},"created_at":"2026-07-05T09:56:51.809455+00:00","updated_at":"2026-07-05T09:56:51.809455+00:00"}