{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DZZFWDJ27NTWYDNJFAL2VZEQJU","short_pith_number":"pith:DZZFWDJ2","schema_version":"1.0","canonical_sha256":"1e725b0d3afb676c0da92817aae4904d09f4145854f79535858d03f2c169a600","source":{"kind":"arxiv","id":"2405.11530","version":1},"attestation_state":"computed","paper":{"title":"Learning More Generalized Experts by Merging Experts in Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Sejik Park","submitted_at":"2024-05-19T11:55:48Z","abstract_excerpt":"We observe that incorporating a shared layer in a mixture-of-experts can lead to performance degradation. This leads us to hypothesize that learning shared features poses challenges in deep learning, potentially caused by the same feature being learned as various different features. To address this issue, we track each expert's usage frequency and merge the two most frequently selected experts. We then update the least frequently selected expert using the combination of experts. This approach, combined with the subsequent learning of the router's expert selection, allows the model to determine"},"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":"2405.11530","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-05-19T11:55:48Z","cross_cats_sorted":[],"title_canon_sha256":"172f1f9d7c9e8a8a1758beaf23d0db7c500ffe79e710b4e6b484c25756001b8d","abstract_canon_sha256":"9d1a2ebc64aaceeb3a8421b5f4f147acd20ee25e45a2a0019e30a232b13dee69"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:20:41.110153Z","signature_b64":"sTXJAdTQuf/pRiW2RnEc8KmI4Jxb1vq9Mb1PcvlcYQeBifFcYHmjXxX3n0l6WKTZKV4O/8ylnoTJA1fJ2AOZBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1e725b0d3afb676c0da92817aae4904d09f4145854f79535858d03f2c169a600","last_reissued_at":"2026-07-05T08:20:41.109639Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:20:41.109639Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning More Generalized Experts by Merging Experts in Mixture-of-Experts","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Sejik Park","submitted_at":"2024-05-19T11:55:48Z","abstract_excerpt":"We observe that incorporating a shared layer in a mixture-of-experts can lead to performance degradation. This leads us to hypothesize that learning shared features poses challenges in deep learning, potentially caused by the same feature being learned as various different features. To address this issue, we track each expert's usage frequency and merge the two most frequently selected experts. We then update the least frequently selected expert using the combination of experts. This approach, combined with the subsequent learning of the router's expert selection, allows the model to determine"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2405.11530","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/2405.11530/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":"2405.11530","created_at":"2026-07-05T08:20:41.109708+00:00"},{"alias_kind":"arxiv_version","alias_value":"2405.11530v1","created_at":"2026-07-05T08:20:41.109708+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2405.11530","created_at":"2026-07-05T08:20:41.109708+00:00"},{"alias_kind":"pith_short_12","alias_value":"DZZFWDJ27NTW","created_at":"2026-07-05T08:20:41.109708+00:00"},{"alias_kind":"pith_short_16","alias_value":"DZZFWDJ27NTWYDNJ","created_at":"2026-07-05T08:20:41.109708+00:00"},{"alias_kind":"pith_short_8","alias_value":"DZZFWDJ2","created_at":"2026-07-05T08:20:41.109708+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.30876","citing_title":"dMoE: dLLMs with Learnable Block Experts","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20408","citing_title":"Spectral Souping: A Unified Framework for Online Preference Alignment","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU","json":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU.json","graph_json":"https://pith.science/api/pith-number/DZZFWDJ27NTWYDNJFAL2VZEQJU/graph.json","events_json":"https://pith.science/api/pith-number/DZZFWDJ27NTWYDNJFAL2VZEQJU/events.json","paper":"https://pith.science/paper/DZZFWDJ2"},"agent_actions":{"view_html":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU","download_json":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU.json","view_paper":"https://pith.science/paper/DZZFWDJ2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2405.11530&json=true","fetch_graph":"https://pith.science/api/pith-number/DZZFWDJ27NTWYDNJFAL2VZEQJU/graph.json","fetch_events":"https://pith.science/api/pith-number/DZZFWDJ27NTWYDNJFAL2VZEQJU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU/action/storage_attestation","attest_author":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU/action/author_attestation","sign_citation":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU/action/citation_signature","submit_replication":"https://pith.science/pith/DZZFWDJ27NTWYDNJFAL2VZEQJU/action/replication_record"}},"created_at":"2026-07-05T08:20:41.109708+00:00","updated_at":"2026-07-05T08:20:41.109708+00:00"}