{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4T4P2Z3KR5NV24XY6QGGX4M76S","short_pith_number":"pith:4T4P2Z3K","schema_version":"1.0","canonical_sha256":"e4f8fd676a8f5b5d72f8f40c6bf19ff49e5462058394083f418ec5197f763924","source":{"kind":"arxiv","id":"2601.18511","version":2},"attestation_state":"computed","paper":{"title":"Scaling up FHE-based Privacy-Preserving ML: Higher Throughput, Longer Inputs for LLama-3-8B","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Damien Stehl\\'e, Guillaume Hanrot, Jai Hyun Park, Jaiyoung Park, Jung Hee Cheon, Jung Ho Ahn, Jung Woo Kim, Minje Park, Sejin Park","submitted_at":"2026-01-26T14:17:23Z","abstract_excerpt":"As large language models (LLMs) become ubiquitous, privacy concerns pertaining to inference keep growing. Fully homomorphic encryption (FHE) has emerged as a primary cryptographic solution for non-interactive confidential LLM inference. However, existing solutions scale poorly with input token length, focusing on small models or input sizes. They also suffer from large outlier values, which strongly impact the evaluation of non-linear layers, leading to heavy polynomial approximation costs. We scale up FHE-based LLM inference in two directions.\n  First, we accelerate FHE-based inference for 12"},"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":"2601.18511","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CR","submitted_at":"2026-01-26T14:17:23Z","cross_cats_sorted":[],"title_canon_sha256":"8d533104b43e42ce325ad9942cb47ed50ce67dd6abb0a43f02ef6d1f2a592f88","abstract_canon_sha256":"4b6c05a30db35b286d8ffebe314f7a887661010be3bea90ef064ca94d92406d1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:11:19.829988Z","signature_b64":"rs5UDar7pY8gJzdiXlrF1owv3KgSD71MtRKxsU3ny1gEFBCBJp/yNe92g7s+Drw1+beUXvO8vUQH0DVuFER8Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e4f8fd676a8f5b5d72f8f40c6bf19ff49e5462058394083f418ec5197f763924","last_reissued_at":"2026-06-19T16:11:19.829539Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:11:19.829539Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Scaling up FHE-based Privacy-Preserving ML: Higher Throughput, Longer Inputs for LLama-3-8B","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Damien Stehl\\'e, Guillaume Hanrot, Jai Hyun Park, Jaiyoung Park, Jung Hee Cheon, Jung Ho Ahn, Jung Woo Kim, Minje Park, Sejin Park","submitted_at":"2026-01-26T14:17:23Z","abstract_excerpt":"As large language models (LLMs) become ubiquitous, privacy concerns pertaining to inference keep growing. Fully homomorphic encryption (FHE) has emerged as a primary cryptographic solution for non-interactive confidential LLM inference. However, existing solutions scale poorly with input token length, focusing on small models or input sizes. They also suffer from large outlier values, which strongly impact the evaluation of non-linear layers, leading to heavy polynomial approximation costs. We scale up FHE-based LLM inference in two directions.\n  First, we accelerate FHE-based inference for 12"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.18511","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/2601.18511/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":"2601.18511","created_at":"2026-06-19T16:11:19.829597+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.18511v2","created_at":"2026-06-19T16:11:19.829597+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.18511","created_at":"2026-06-19T16:11:19.829597+00:00"},{"alias_kind":"pith_short_12","alias_value":"4T4P2Z3KR5NV","created_at":"2026-06-19T16:11:19.829597+00:00"},{"alias_kind":"pith_short_16","alias_value":"4T4P2Z3KR5NV24XY","created_at":"2026-06-19T16:11:19.829597+00:00"},{"alias_kind":"pith_short_8","alias_value":"4T4P2Z3K","created_at":"2026-06-19T16:11:19.829597+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S","json":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S.json","graph_json":"https://pith.science/api/pith-number/4T4P2Z3KR5NV24XY6QGGX4M76S/graph.json","events_json":"https://pith.science/api/pith-number/4T4P2Z3KR5NV24XY6QGGX4M76S/events.json","paper":"https://pith.science/paper/4T4P2Z3K"},"agent_actions":{"view_html":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S","download_json":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S.json","view_paper":"https://pith.science/paper/4T4P2Z3K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.18511&json=true","fetch_graph":"https://pith.science/api/pith-number/4T4P2Z3KR5NV24XY6QGGX4M76S/graph.json","fetch_events":"https://pith.science/api/pith-number/4T4P2Z3KR5NV24XY6QGGX4M76S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S/action/storage_attestation","attest_author":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S/action/author_attestation","sign_citation":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S/action/citation_signature","submit_replication":"https://pith.science/pith/4T4P2Z3KR5NV24XY6QGGX4M76S/action/replication_record"}},"created_at":"2026-06-19T16:11:19.829597+00:00","updated_at":"2026-06-19T16:11:19.829597+00:00"}