{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:QBGHUMTFGFPPOI5GF3ROCHAIJG","short_pith_number":"pith:QBGHUMTF","schema_version":"1.0","canonical_sha256":"804c7a3265315ef723a62ee2e11c08498343df0402717fd9ac1d0de23996b309","source":{"kind":"arxiv","id":"2506.12597","version":1},"attestation_state":"computed","paper":{"title":"Automatic Expert Discovery in LLM Upcycling via Sparse Interpolated Mixture-of-Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jonathan Richard Schwarz, Shengzhuang Chen, Ying Wei","submitted_at":"2025-06-14T18:34:38Z","abstract_excerpt":"We present Sparse Interpolated Mixture-of-Experts (SIMoE) instruction-tuning, an end-to-end algorithm designed to fine-tune a dense pre-trained Large Language Model (LLM) into a MoE-style model that possesses capabilities in multiple specialized domains. During instruction-tuning, SIMoE automatically identifies multiple specialized experts under a specified sparsity constraint, with each expert representing a structurally sparse subset of the seed LLM's parameters that correspond to domain-specific knowledge within the data. SIMoE simultaneously learns an input-dependent expert merging strateg"},"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":"2506.12597","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-14T18:34:38Z","cross_cats_sorted":[],"title_canon_sha256":"3ea1ef4bba226ee8771c8100118e4405427394277b7e06620461cca4b2e91574","abstract_canon_sha256":"2a0285dc989f78ced5aaa3d451a6e284090517d0e775efc290dadf5f87c9788c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:56.119089Z","signature_b64":"u9WDeUS8KNxwu1jHThFJpo42dFdr+poMbA3WTxcanSjoKS4pQ0nZiwnaaD3YCgYaq8Tqnp3uv/5lQ6ZaivbzCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"804c7a3265315ef723a62ee2e11c08498343df0402717fd9ac1d0de23996b309","last_reissued_at":"2026-07-05T11:21:56.118621Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:56.118621Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Automatic Expert Discovery in LLM Upcycling via Sparse Interpolated Mixture-of-Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Jonathan Richard Schwarz, Shengzhuang Chen, Ying Wei","submitted_at":"2025-06-14T18:34:38Z","abstract_excerpt":"We present Sparse Interpolated Mixture-of-Experts (SIMoE) instruction-tuning, an end-to-end algorithm designed to fine-tune a dense pre-trained Large Language Model (LLM) into a MoE-style model that possesses capabilities in multiple specialized domains. During instruction-tuning, SIMoE automatically identifies multiple specialized experts under a specified sparsity constraint, with each expert representing a structurally sparse subset of the seed LLM's parameters that correspond to domain-specific knowledge within the data. SIMoE simultaneously learns an input-dependent expert merging strateg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.12597","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/2506.12597/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":"2506.12597","created_at":"2026-07-05T11:21:56.118677+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.12597v1","created_at":"2026-07-05T11:21:56.118677+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.12597","created_at":"2026-07-05T11:21:56.118677+00:00"},{"alias_kind":"pith_short_12","alias_value":"QBGHUMTFGFPP","created_at":"2026-07-05T11:21:56.118677+00:00"},{"alias_kind":"pith_short_16","alias_value":"QBGHUMTFGFPPOI5G","created_at":"2026-07-05T11:21:56.118677+00:00"},{"alias_kind":"pith_short_8","alias_value":"QBGHUMTF","created_at":"2026-07-05T11:21:56.118677+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/QBGHUMTFGFPPOI5GF3ROCHAIJG","json":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG.json","graph_json":"https://pith.science/api/pith-number/QBGHUMTFGFPPOI5GF3ROCHAIJG/graph.json","events_json":"https://pith.science/api/pith-number/QBGHUMTFGFPPOI5GF3ROCHAIJG/events.json","paper":"https://pith.science/paper/QBGHUMTF"},"agent_actions":{"view_html":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG","download_json":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG.json","view_paper":"https://pith.science/paper/QBGHUMTF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.12597&json=true","fetch_graph":"https://pith.science/api/pith-number/QBGHUMTFGFPPOI5GF3ROCHAIJG/graph.json","fetch_events":"https://pith.science/api/pith-number/QBGHUMTFGFPPOI5GF3ROCHAIJG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG/action/storage_attestation","attest_author":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG/action/author_attestation","sign_citation":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG/action/citation_signature","submit_replication":"https://pith.science/pith/QBGHUMTFGFPPOI5GF3ROCHAIJG/action/replication_record"}},"created_at":"2026-07-05T11:21:56.118677+00:00","updated_at":"2026-07-05T11:21:56.118677+00:00"}