{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:X2QVYOFPI22MFXZMYIIL2FWGVD","short_pith_number":"pith:X2QVYOFP","schema_version":"1.0","canonical_sha256":"bea15c38af46b4c2df2cc210bd16c6a8e3a4acab80fcb0caab96a9a56e7aa3a6","source":{"kind":"arxiv","id":"2406.12034","version":2},"attestation_state":"computed","paper":{"title":"Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alan Ritter, David Cox, Hongyin Luo, Jacob Hansen, James Glass, Junmo Kang, Leonid Karlinsky, Rameswar Panda, Rogerio Feris, Zhen Wang","submitted_at":"2024-06-17T19:06:54Z","abstract_excerpt":"We present Self-MoE, an approach that transforms a monolithic LLM into a compositional, modular system of self-specialized experts, named MiXSE (MiXture of Self-specialized Experts). Our approach leverages self-specialization, which constructs expert modules using self-generated synthetic data, each equipping a shared base LLM with distinct domain-specific capabilities, activated via self-optimized routing. This allows for dynamic and capability-specific handling of various target tasks, enhancing overall capabilities, without extensive human-labeled data and added parameters. Our empirical re"},"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":"2406.12034","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-17T19:06:54Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"4c5fb3423c3f13a9b1550acd5cc198785b643e7d933effa90744bda202673d3e","abstract_canon_sha256":"58b886d612f0e132d1c86ab7f578201abd87c8842e99bac0ec2900ee58ba3892"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:16:50.228670Z","signature_b64":"Fs+be54diAhy4elUKwg12pB9ngGnDDHYQ8tXU+xP6aIjPvHMXWOmb8b9TdQQfgv46BLQhp1Et/yWVlonnlpWCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bea15c38af46b4c2df2cc210bd16c6a8e3a4acab80fcb0caab96a9a56e7aa3a6","last_reissued_at":"2026-07-05T09:16:50.228167Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:16:50.228167Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Alan Ritter, David Cox, Hongyin Luo, Jacob Hansen, James Glass, Junmo Kang, Leonid Karlinsky, Rameswar Panda, Rogerio Feris, Zhen Wang","submitted_at":"2024-06-17T19:06:54Z","abstract_excerpt":"We present Self-MoE, an approach that transforms a monolithic LLM into a compositional, modular system of self-specialized experts, named MiXSE (MiXture of Self-specialized Experts). Our approach leverages self-specialization, which constructs expert modules using self-generated synthetic data, each equipping a shared base LLM with distinct domain-specific capabilities, activated via self-optimized routing. This allows for dynamic and capability-specific handling of various target tasks, enhancing overall capabilities, without extensive human-labeled data and added parameters. Our empirical re"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.12034","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/2406.12034/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":"2406.12034","created_at":"2026-07-05T09:16:50.228227+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.12034v2","created_at":"2026-07-05T09:16:50.228227+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.12034","created_at":"2026-07-05T09:16:50.228227+00:00"},{"alias_kind":"pith_short_12","alias_value":"X2QVYOFPI22M","created_at":"2026-07-05T09:16:50.228227+00:00"},{"alias_kind":"pith_short_16","alias_value":"X2QVYOFPI22MFXZM","created_at":"2026-07-05T09:16:50.228227+00:00"},{"alias_kind":"pith_short_8","alias_value":"X2QVYOFP","created_at":"2026-07-05T09:16:50.228227+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2408.07666","citing_title":"Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities","ref_index":107,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12995","citing_title":"PolicyLLM: Towards Excellent Comprehension of Public Policy for Large Language Models","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD","json":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD.json","graph_json":"https://pith.science/api/pith-number/X2QVYOFPI22MFXZMYIIL2FWGVD/graph.json","events_json":"https://pith.science/api/pith-number/X2QVYOFPI22MFXZMYIIL2FWGVD/events.json","paper":"https://pith.science/paper/X2QVYOFP"},"agent_actions":{"view_html":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD","download_json":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD.json","view_paper":"https://pith.science/paper/X2QVYOFP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.12034&json=true","fetch_graph":"https://pith.science/api/pith-number/X2QVYOFPI22MFXZMYIIL2FWGVD/graph.json","fetch_events":"https://pith.science/api/pith-number/X2QVYOFPI22MFXZMYIIL2FWGVD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD/action/storage_attestation","attest_author":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD/action/author_attestation","sign_citation":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD/action/citation_signature","submit_replication":"https://pith.science/pith/X2QVYOFPI22MFXZMYIIL2FWGVD/action/replication_record"}},"created_at":"2026-07-05T09:16:50.228227+00:00","updated_at":"2026-07-05T09:16:50.228227+00:00"}