{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BW72VCTWHOBTGKCHXZ24AMUXWD","short_pith_number":"pith:BW72VCTW","schema_version":"1.0","canonical_sha256":"0dbfaa8a763b83332847be75c03297b0c16a31fcb1332317374d441944700664","source":{"kind":"arxiv","id":"2406.13233","version":2},"attestation_state":"computed","paper":{"title":"AdaMoE: Token-Adaptive Routing with Null Experts for Mixture-of-Experts Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Hao Zhang, Hongcheng Gao, Yibo Miao, Zhijie Deng, Zihao Zeng","submitted_at":"2024-06-19T05:47:10Z","abstract_excerpt":"Mixture of experts (MoE) has become the standard for constructing production-level large language models (LLMs) due to its promise to boost model capacity without causing significant overheads. Nevertheless, existing MoE methods usually enforce a constant top-k routing for all tokens, which is arguably restrictive because various tokens (e.g., \"<EOS>\" vs. \"apple\") may require various numbers of experts for feature abstraction. Lifting such a constraint can help make the most of limited resources and unleash the potential of the model for downstream tasks. In this sense, we introduce AdaMoE to "},"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":"2406.13233","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-06-19T05:47:10Z","cross_cats_sorted":[],"title_canon_sha256":"6f308517723a2a7629e43ccbf415358fc6e8707ef455029d8ce29500409d56cc","abstract_canon_sha256":"e9561ccade0a66016a4eac719fb6021956eca09823879419048a28d686b844a4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:59.656838Z","signature_b64":"irkxh50ZDP8hs1XeAxgPkFagB4opoumgZCzUp6l+Z2ADgZFw0gwxXXSMYIZiLvZNKvbLDI7qkvDHmEXmZ3nsCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0dbfaa8a763b83332847be75c03297b0c16a31fcb1332317374d441944700664","last_reissued_at":"2026-07-05T09:19:59.656334Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:59.656334Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AdaMoE: Token-Adaptive Routing with Null Experts for Mixture-of-Experts Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Hao Zhang, Hongcheng Gao, Yibo Miao, Zhijie Deng, Zihao Zeng","submitted_at":"2024-06-19T05:47:10Z","abstract_excerpt":"Mixture of experts (MoE) has become the standard for constructing production-level large language models (LLMs) due to its promise to boost model capacity without causing significant overheads. Nevertheless, existing MoE methods usually enforce a constant top-k routing for all tokens, which is arguably restrictive because various tokens (e.g., \"<EOS>\" vs. \"apple\") may require various numbers of experts for feature abstraction. Lifting such a constraint can help make the most of limited resources and unleash the potential of the model for downstream tasks. In this sense, we introduce AdaMoE to "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.13233","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/2406.13233/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":"2406.13233","created_at":"2026-07-05T09:19:59.656404+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.13233v2","created_at":"2026-07-05T09:19:59.656404+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.13233","created_at":"2026-07-05T09:19:59.656404+00:00"},{"alias_kind":"pith_short_12","alias_value":"BW72VCTWHOBT","created_at":"2026-07-05T09:19:59.656404+00:00"},{"alias_kind":"pith_short_16","alias_value":"BW72VCTWHOBTGKCH","created_at":"2026-07-05T09:19:59.656404+00:00"},{"alias_kind":"pith_short_8","alias_value":"BW72VCTW","created_at":"2026-07-05T09:19:59.656404+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.07785","citing_title":"Grove MoE: Towards Efficient and Superior MoE LLMs with Adjugate Experts","ref_index":49,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD","json":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD.json","graph_json":"https://pith.science/api/pith-number/BW72VCTWHOBTGKCHXZ24AMUXWD/graph.json","events_json":"https://pith.science/api/pith-number/BW72VCTWHOBTGKCHXZ24AMUXWD/events.json","paper":"https://pith.science/paper/BW72VCTW"},"agent_actions":{"view_html":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD","download_json":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD.json","view_paper":"https://pith.science/paper/BW72VCTW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.13233&json=true","fetch_graph":"https://pith.science/api/pith-number/BW72VCTWHOBTGKCHXZ24AMUXWD/graph.json","fetch_events":"https://pith.science/api/pith-number/BW72VCTWHOBTGKCHXZ24AMUXWD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD/action/storage_attestation","attest_author":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD/action/author_attestation","sign_citation":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD/action/citation_signature","submit_replication":"https://pith.science/pith/BW72VCTWHOBTGKCHXZ24AMUXWD/action/replication_record"}},"created_at":"2026-07-05T09:19:59.656404+00:00","updated_at":"2026-07-05T09:19:59.656404+00:00"}