{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NE75VE5J5SC6IM7FNZNIJAYOW2","short_pith_number":"pith:NE75VE5J","schema_version":"1.0","canonical_sha256":"693fda93a9ec85e433e56e5a84830eb695b89da779cfe2d392bbf0f35a08655d","source":{"kind":"arxiv","id":"2608.01784","version":1},"attestation_state":"computed","paper":{"title":"REFLEX: Rethinking MoE Inference as Refinement-Aware Compute Allocation in Diffusion Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Cheng Yan, Hongyu Zhang, Jiazheng Liu, Wuyang Zhang, Xiang Xia, Yiming Zhang","submitted_at":"2026-08-03T06:59:06Z","abstract_excerpt":"Mixture-of-experts (MoE) models increase parameter capacity by activating only a small subset of experts for each token. This conditional-computation paradigm has enabled autoregressive language models to scale model capacity without a proportional increase in per-token computation. In diffusion language models (DLMs), however, each denoising forward jointly revisits all token positions despite their sharply different refinement demands, while the default fixed token-choice routing assigns them a uniform expert budget, creating a mismatch between expert computation and refinement demand. We ar"},"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":"2608.01784","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-03T06:59:06Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"17ff78b7a8ff9e1d27563b45675c1aa8b6b92bd7e67d8db1788ee6596946fc30","abstract_canon_sha256":"b2b89873df12a1a4b4c244b868ec1df7b083a10981178f804c9479649514dc12"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:07:04.930427Z","signature_b64":"N8yyDmojPApuOncGKqWhu0T0Zl8qV169UaNUyYJDMrDRWISptFlpWqIT18A6aUYaZp1wGfYCSA1PYlbkQDO2Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"693fda93a9ec85e433e56e5a84830eb695b89da779cfe2d392bbf0f35a08655d","last_reissued_at":"2026-08-04T02:07:04.928940Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:07:04.928940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"REFLEX: Rethinking MoE Inference as Refinement-Aware Compute Allocation in Diffusion Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Cheng Yan, Hongyu Zhang, Jiazheng Liu, Wuyang Zhang, Xiang Xia, Yiming Zhang","submitted_at":"2026-08-03T06:59:06Z","abstract_excerpt":"Mixture-of-experts (MoE) models increase parameter capacity by activating only a small subset of experts for each token. This conditional-computation paradigm has enabled autoregressive language models to scale model capacity without a proportional increase in per-token computation. In diffusion language models (DLMs), however, each denoising forward jointly revisits all token positions despite their sharply different refinement demands, while the default fixed token-choice routing assigns them a uniform expert budget, creating a mismatch between expert computation and refinement demand. We ar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01784","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/2608.01784/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":"2608.01784","created_at":"2026-08-04T02:07:04.930386+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.01784v1","created_at":"2026-08-04T02:07:04.930386+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01784","created_at":"2026-08-04T02:07:04.930386+00:00"},{"alias_kind":"pith_short_12","alias_value":"NE75VE5J5SC6","created_at":"2026-08-04T02:07:04.930386+00:00"},{"alias_kind":"pith_short_16","alias_value":"NE75VE5J5SC6IM7F","created_at":"2026-08-04T02:07:04.930386+00:00"},{"alias_kind":"pith_short_8","alias_value":"NE75VE5J","created_at":"2026-08-04T02:07:04.930386+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/NE75VE5J5SC6IM7FNZNIJAYOW2","json":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2.json","graph_json":"https://pith.science/api/pith-number/NE75VE5J5SC6IM7FNZNIJAYOW2/graph.json","events_json":"https://pith.science/api/pith-number/NE75VE5J5SC6IM7FNZNIJAYOW2/events.json","paper":"https://pith.science/paper/NE75VE5J"},"agent_actions":{"view_html":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2","download_json":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2.json","view_paper":"https://pith.science/paper/NE75VE5J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.01784&json=true","fetch_graph":"https://pith.science/api/pith-number/NE75VE5J5SC6IM7FNZNIJAYOW2/graph.json","fetch_events":"https://pith.science/api/pith-number/NE75VE5J5SC6IM7FNZNIJAYOW2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2/action/storage_attestation","attest_author":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2/action/author_attestation","sign_citation":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2/action/citation_signature","submit_replication":"https://pith.science/pith/NE75VE5J5SC6IM7FNZNIJAYOW2/action/replication_record"}},"created_at":"2026-08-04T02:07:04.930386+00:00","updated_at":"2026-08-04T02:07:04.930386+00:00"}