{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FIT5LHXXS32CGASBVD2PJ3ASXX","short_pith_number":"pith:FIT5LHXX","schema_version":"1.0","canonical_sha256":"2a27d59ef796f4230241a8f4f4ec12bdccf4d7fcd00b77621bdf82c028b46046","source":{"kind":"arxiv","id":"2605.18498","version":1},"attestation_state":"computed","paper":{"title":"DBES: A Systematic Benchmark and Metric Suite for Evaluating Expert Specialization in Large-Scale MoEs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hongxuan Lu, Jazze Young, Jing Wang, Shu Wang, Zhimin Xin","submitted_at":"2026-05-18T14:50:30Z","abstract_excerpt":"Expert specialization in Mixture-of-Experts (MoE) models remains poorly understood, with traditional evaluations conflating architectural load-balancing with functional specialization. We introduce DBES, a comprehensive diagnostic framework combining a multi-domain benchmark with five theoretically grounded metrics: Routing Specialization, Normalized Effective Rank, Domain Isolation, Routing Stiffness Score, and N-gram Expertise measures.\n  Critical findings demonstrate distinct specialization paradigms across models: Qwen-series exhibit modular specialization with high domain isolation, while"},"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":"2605.18498","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-18T14:50:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"d328e62a17570540bfbefd4683bd294bee8a7dbd637f817d4ada182fd19989df","abstract_canon_sha256":"19a4f18180e5bb6baa4b0c510fdf1ed76ba4f99b08cc41d205e8d19f057f082e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:06:04.337190Z","signature_b64":"EdjzBH7DN9wV6bU8PV4aE5zZnMB5DaaM1lEj9EK1LiBeidxWZY9IWJbUiNLoL1VxdXE5dQLiXYdgWcxLkCvoDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2a27d59ef796f4230241a8f4f4ec12bdccf4d7fcd00b77621bdf82c028b46046","last_reissued_at":"2026-05-20T00:06:04.336549Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:06:04.336549Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DBES: A Systematic Benchmark and Metric Suite for Evaluating Expert Specialization in Large-Scale MoEs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hongxuan Lu, Jazze Young, Jing Wang, Shu Wang, Zhimin Xin","submitted_at":"2026-05-18T14:50:30Z","abstract_excerpt":"Expert specialization in Mixture-of-Experts (MoE) models remains poorly understood, with traditional evaluations conflating architectural load-balancing with functional specialization. We introduce DBES, a comprehensive diagnostic framework combining a multi-domain benchmark with five theoretically grounded metrics: Routing Specialization, Normalized Effective Rank, Domain Isolation, Routing Stiffness Score, and N-gram Expertise measures.\n  Critical findings demonstrate distinct specialization paradigms across models: Qwen-series exhibit modular specialization with high domain isolation, while"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18498","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/2605.18498/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":"2605.18498","created_at":"2026-05-20T00:06:04.336633+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18498v1","created_at":"2026-05-20T00:06:04.336633+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18498","created_at":"2026-05-20T00:06:04.336633+00:00"},{"alias_kind":"pith_short_12","alias_value":"FIT5LHXXS32C","created_at":"2026-05-20T00:06:04.336633+00:00"},{"alias_kind":"pith_short_16","alias_value":"FIT5LHXXS32CGASB","created_at":"2026-05-20T00:06:04.336633+00:00"},{"alias_kind":"pith_short_8","alias_value":"FIT5LHXX","created_at":"2026-05-20T00:06:04.336633+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/FIT5LHXXS32CGASBVD2PJ3ASXX","json":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX.json","graph_json":"https://pith.science/api/pith-number/FIT5LHXXS32CGASBVD2PJ3ASXX/graph.json","events_json":"https://pith.science/api/pith-number/FIT5LHXXS32CGASBVD2PJ3ASXX/events.json","paper":"https://pith.science/paper/FIT5LHXX"},"agent_actions":{"view_html":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX","download_json":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX.json","view_paper":"https://pith.science/paper/FIT5LHXX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18498&json=true","fetch_graph":"https://pith.science/api/pith-number/FIT5LHXXS32CGASBVD2PJ3ASXX/graph.json","fetch_events":"https://pith.science/api/pith-number/FIT5LHXXS32CGASBVD2PJ3ASXX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX/action/storage_attestation","attest_author":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX/action/author_attestation","sign_citation":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX/action/citation_signature","submit_replication":"https://pith.science/pith/FIT5LHXXS32CGASBVD2PJ3ASXX/action/replication_record"}},"created_at":"2026-05-20T00:06:04.336633+00:00","updated_at":"2026-05-20T00:06:04.336633+00:00"}