{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:VLW4GSPJTXXDNYQUFLLF7C6SXH","short_pith_number":"pith:VLW4GSPJ","schema_version":"1.0","canonical_sha256":"aaedc349e99dee36e2142ad65f8bd2b9cdf4c048d38e1ee706d8763f92fc49ef","source":{"kind":"arxiv","id":"2401.09796","version":2},"attestation_state":"computed","paper":{"title":"A Fast, Performant, Secure Distributed Training Framework For Large Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Aihui Zhou, Anda Cheng, Chaofan Yu, Lei Wang, Wei Huang, Yinggui Wang","submitted_at":"2024-01-18T08:33:09Z","abstract_excerpt":"The distributed (federated) LLM is an important method for co-training the domain-specific LLM using siloed data. However, maliciously stealing model parameters and data from the server or client side has become an urgent problem to be solved. In this paper, we propose a secure distributed LLM based on model slicing. In this case, we deploy the Trusted Execution Environment (TEE) on both the client and server side, and put the fine-tuned structure (LoRA or embedding of P-tuning v2) into the TEE. Then, secure communication is executed in the TEE and general environments through lightweight encr"},"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":"2401.09796","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-01-18T08:33:09Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"7e95ebdda2947c29cb278ebfe9d07bfc6d3c065dda7342f25b3797d5e0b7eae7","abstract_canon_sha256":"b44e294f1831082ca358cf4ffe08444585c23d0ece553626311f5f4e855c0572"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:35:24.683423Z","signature_b64":"PNfnQZ0Bze9oqgh7GLH/mdX+27QAMg1wsYk0xFUzQ/PEug/5glSKO3Z7qq9USduBz77Mg47gnahve151juQLAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"aaedc349e99dee36e2142ad65f8bd2b9cdf4c048d38e1ee706d8763f92fc49ef","last_reissued_at":"2026-07-05T07:35:24.683003Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:35:24.683003Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Fast, Performant, Secure Distributed Training Framework For Large Language Model","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.LG","authors_text":"Aihui Zhou, Anda Cheng, Chaofan Yu, Lei Wang, Wei Huang, Yinggui Wang","submitted_at":"2024-01-18T08:33:09Z","abstract_excerpt":"The distributed (federated) LLM is an important method for co-training the domain-specific LLM using siloed data. However, maliciously stealing model parameters and data from the server or client side has become an urgent problem to be solved. In this paper, we propose a secure distributed LLM based on model slicing. In this case, we deploy the Trusted Execution Environment (TEE) on both the client and server side, and put the fine-tuned structure (LoRA or embedding of P-tuning v2) into the TEE. Then, secure communication is executed in the TEE and general environments through lightweight encr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.09796","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/2401.09796/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":"2401.09796","created_at":"2026-07-05T07:35:24.683059+00:00"},{"alias_kind":"arxiv_version","alias_value":"2401.09796v2","created_at":"2026-07-05T07:35:24.683059+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.09796","created_at":"2026-07-05T07:35:24.683059+00:00"},{"alias_kind":"pith_short_12","alias_value":"VLW4GSPJTXXD","created_at":"2026-07-05T07:35:24.683059+00:00"},{"alias_kind":"pith_short_16","alias_value":"VLW4GSPJTXXDNYQU","created_at":"2026-07-05T07:35:24.683059+00:00"},{"alias_kind":"pith_short_8","alias_value":"VLW4GSPJ","created_at":"2026-07-05T07:35:24.683059+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.01976","citing_title":"A Survey on Privacy Risks and Protection in Large Language Models","ref_index":74,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH","json":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH.json","graph_json":"https://pith.science/api/pith-number/VLW4GSPJTXXDNYQUFLLF7C6SXH/graph.json","events_json":"https://pith.science/api/pith-number/VLW4GSPJTXXDNYQUFLLF7C6SXH/events.json","paper":"https://pith.science/paper/VLW4GSPJ"},"agent_actions":{"view_html":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH","download_json":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH.json","view_paper":"https://pith.science/paper/VLW4GSPJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2401.09796&json=true","fetch_graph":"https://pith.science/api/pith-number/VLW4GSPJTXXDNYQUFLLF7C6SXH/graph.json","fetch_events":"https://pith.science/api/pith-number/VLW4GSPJTXXDNYQUFLLF7C6SXH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH/action/storage_attestation","attest_author":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH/action/author_attestation","sign_citation":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH/action/citation_signature","submit_replication":"https://pith.science/pith/VLW4GSPJTXXDNYQUFLLF7C6SXH/action/replication_record"}},"created_at":"2026-07-05T07:35:24.683059+00:00","updated_at":"2026-07-05T07:35:24.683059+00:00"}