{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:3QFLE6C6VAA44X7KLU3KLUGEBP","short_pith_number":"pith:3QFLE6C6","schema_version":"1.0","canonical_sha256":"dc0ab2785ea801ce5fea5d36a5d0c40bc08a0738a21e3c5a7e5b7ba6a7e26a24","source":{"kind":"arxiv","id":"2508.01261","version":1},"attestation_state":"computed","paper":{"title":"Unifying Mixture of Experts and Multi-Head Latent Attention for Efficient Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Rajat Dandekar, Raj Dandekar, Sreedath Panat, Sushant Mehta","submitted_at":"2025-08-02T08:33:30Z","abstract_excerpt":"We present MoE-MLA-RoPE, a novel architecture combination that combines Mixture of Experts (MoE) with Multi-head Latent Attention (MLA) and Rotary Position Embeddings (RoPE) for efficient language modeling. Our approach addresses the fundamental trade-off between model capacity and computational efficiency through three key innovations: (1) fine-grained expert routing with 64 micro-experts and top-$k$ selection, enabling flexible specialization through 3.6 * 10^7 possible expert combinations; (2) shared expert isolation that dedicates 2 always active experts for common patterns while routing t"},"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":"2508.01261","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-08-02T08:33:30Z","cross_cats_sorted":[],"title_canon_sha256":"657b05fc2b6de63b299d55d793be72f031c132e284d7ea78e2aa0f2c9c947cc5","abstract_canon_sha256":"e6dd0717e7c4a8813bbc82df3e1eeedbd8dbc524eaa1c8ffa6f33a9d83cca9b4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:47:36.149291Z","signature_b64":"g5vpIVNyWg1wHcH2YaYIwe0854KLN9pzxzQqpOasXU0rlnfUJ2ya+dZLB7Ix07Vsqm+gAIId6qb1giJ7MefjBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dc0ab2785ea801ce5fea5d36a5d0c40bc08a0738a21e3c5a7e5b7ba6a7e26a24","last_reissued_at":"2026-07-05T11:47:36.148831Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:47:36.148831Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unifying Mixture of Experts and Multi-Head Latent Attention for Efficient Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Rajat Dandekar, Raj Dandekar, Sreedath Panat, Sushant Mehta","submitted_at":"2025-08-02T08:33:30Z","abstract_excerpt":"We present MoE-MLA-RoPE, a novel architecture combination that combines Mixture of Experts (MoE) with Multi-head Latent Attention (MLA) and Rotary Position Embeddings (RoPE) for efficient language modeling. Our approach addresses the fundamental trade-off between model capacity and computational efficiency through three key innovations: (1) fine-grained expert routing with 64 micro-experts and top-$k$ selection, enabling flexible specialization through 3.6 * 10^7 possible expert combinations; (2) shared expert isolation that dedicates 2 always active experts for common patterns while routing t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.01261","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/2508.01261/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":"2508.01261","created_at":"2026-07-05T11:47:36.148891+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.01261v1","created_at":"2026-07-05T11:47:36.148891+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.01261","created_at":"2026-07-05T11:47:36.148891+00:00"},{"alias_kind":"pith_short_12","alias_value":"3QFLE6C6VAA4","created_at":"2026-07-05T11:47:36.148891+00:00"},{"alias_kind":"pith_short_16","alias_value":"3QFLE6C6VAA44X7K","created_at":"2026-07-05T11:47:36.148891+00:00"},{"alias_kind":"pith_short_8","alias_value":"3QFLE6C6","created_at":"2026-07-05T11:47:36.148891+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/3QFLE6C6VAA44X7KLU3KLUGEBP","json":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP.json","graph_json":"https://pith.science/api/pith-number/3QFLE6C6VAA44X7KLU3KLUGEBP/graph.json","events_json":"https://pith.science/api/pith-number/3QFLE6C6VAA44X7KLU3KLUGEBP/events.json","paper":"https://pith.science/paper/3QFLE6C6"},"agent_actions":{"view_html":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP","download_json":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP.json","view_paper":"https://pith.science/paper/3QFLE6C6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.01261&json=true","fetch_graph":"https://pith.science/api/pith-number/3QFLE6C6VAA44X7KLU3KLUGEBP/graph.json","fetch_events":"https://pith.science/api/pith-number/3QFLE6C6VAA44X7KLU3KLUGEBP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP/action/storage_attestation","attest_author":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP/action/author_attestation","sign_citation":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP/action/citation_signature","submit_replication":"https://pith.science/pith/3QFLE6C6VAA44X7KLU3KLUGEBP/action/replication_record"}},"created_at":"2026-07-05T11:47:36.148891+00:00","updated_at":"2026-07-05T11:47:36.148891+00:00"}