{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:DVQ5T4IDYZOFT3GJFWVWDM2ZH3","short_pith_number":"pith:DVQ5T4ID","schema_version":"1.0","canonical_sha256":"1d61d9f103c65c59ecc92dab61b3593eebe73039e2cafc64badcb4d0a3bd951f","source":{"kind":"arxiv","id":"2503.06491","version":1},"attestation_state":"computed","paper":{"title":"MoFE: Mixture of Frozen Experts Architecture","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Hyopil Shin, Jaeyoon Kim, Jean Seo","submitted_at":"2025-03-09T07:24:36Z","abstract_excerpt":"We propose the Mixture of Frozen Experts (MoFE) architecture, which integrates Parameter-efficient Fine-tuning (PEFT) and the Mixture of Experts (MoE) architecture to enhance both training efficiency and model scalability. By freezing the Feed Forward Network (FFN) layers within the MoE framework, MoFE significantly reduces the number of trainable parameters, improving training efficiency while still allowing for effective knowledge transfer from the expert models. This facilitates the creation of models proficient in multiple domains. We conduct experiments to evaluate the trade-offs between "},"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":"2503.06491","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-09T07:24:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8fd9c3102f22825e8bfa0c7c81e2a31466bcb4cf2bda16cb407f791a6b77a105","abstract_canon_sha256":"6b895d88ccb8e79cb0a59388e01d1c175b01e6d5dcfe4d16b8941a439c650908"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:35.201993Z","signature_b64":"sKJ7QWV3YMIsVwgK6qr5l+qKGP1vXybIm7tsDd0bwQo0js7L6W48Q8CWz5pbTlimTDLfPDWkg99IrlZrLkt3Bw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1d61d9f103c65c59ecc92dab61b3593eebe73039e2cafc64badcb4d0a3bd951f","last_reissued_at":"2026-07-05T10:27:35.201502Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:35.201502Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MoFE: Mixture of Frozen Experts Architecture","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Hyopil Shin, Jaeyoon Kim, Jean Seo","submitted_at":"2025-03-09T07:24:36Z","abstract_excerpt":"We propose the Mixture of Frozen Experts (MoFE) architecture, which integrates Parameter-efficient Fine-tuning (PEFT) and the Mixture of Experts (MoE) architecture to enhance both training efficiency and model scalability. By freezing the Feed Forward Network (FFN) layers within the MoE framework, MoFE significantly reduces the number of trainable parameters, improving training efficiency while still allowing for effective knowledge transfer from the expert models. This facilitates the creation of models proficient in multiple domains. We conduct experiments to evaluate the trade-offs between "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06491","kind":"arxiv","version":1},"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/2503.06491/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":"2503.06491","created_at":"2026-07-05T10:27:35.201562+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.06491v1","created_at":"2026-07-05T10:27:35.201562+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06491","created_at":"2026-07-05T10:27:35.201562+00:00"},{"alias_kind":"pith_short_12","alias_value":"DVQ5T4IDYZOF","created_at":"2026-07-05T10:27:35.201562+00:00"},{"alias_kind":"pith_short_16","alias_value":"DVQ5T4IDYZOFT3GJ","created_at":"2026-07-05T10:27:35.201562+00:00"},{"alias_kind":"pith_short_8","alias_value":"DVQ5T4ID","created_at":"2026-07-05T10:27:35.201562+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.22919","citing_title":"Hecto: Modular Sparse Experts for Adaptive and Interpretable Reasoning","ref_index":9,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3","json":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3.json","graph_json":"https://pith.science/api/pith-number/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/graph.json","events_json":"https://pith.science/api/pith-number/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/events.json","paper":"https://pith.science/paper/DVQ5T4ID"},"agent_actions":{"view_html":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3","download_json":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3.json","view_paper":"https://pith.science/paper/DVQ5T4ID","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.06491&json=true","fetch_graph":"https://pith.science/api/pith-number/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/graph.json","fetch_events":"https://pith.science/api/pith-number/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/action/storage_attestation","attest_author":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/action/author_attestation","sign_citation":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/action/citation_signature","submit_replication":"https://pith.science/pith/DVQ5T4IDYZOFT3GJFWVWDM2ZH3/action/replication_record"}},"created_at":"2026-07-05T10:27:35.201562+00:00","updated_at":"2026-07-05T10:27:35.201562+00:00"}