{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KIDMCY2PBTH7HGKAZU5AHQMFBJ","short_pith_number":"pith:KIDMCY2P","schema_version":"1.0","canonical_sha256":"5206c1634f0ccff39940cd3a03c1850a72507de903e0053a12fbd07d11499e8a","source":{"kind":"arxiv","id":"2607.06601","version":1},"attestation_state":"computed","paper":{"title":"TriRoute: Unified Learned Routing for Joint Adaptive Attention, Experts, and KV-Cache Allocation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andrii Balashov, Olena Ponomarova","submitted_at":"2026-07-07T00:12:46Z","abstract_excerpt":"Conditional computation can decouple language model quality from per-token inference cost, yet leading techniques act on a single axis in isolation: Mixture-of-Experts (MoE) sparsifies the FFN, Mixture-of-Depths (MoD) skips whole transformer blocks, and KV-cache quantization compresses attention memory. We argue these three decisions (attention resolution, expert selection, and cache bit-width) are strongly coupled and should be made jointly: a token rare enough to warrant full attention may also need high-precision caching regardless of which expert processes it. We introduce TriRoute, a sing"},"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":"2607.06601","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-07T00:12:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4abf3ec9d092c0681b59c31b5d9e89b1434ead4c5ffd4d083897b5442ca16025","abstract_canon_sha256":"5089c8b02809d98cd7bffbbec4c130e095a11d24f6ef30b11285c313d46515cf"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-09T00:19:19.896169Z","signature_b64":"upybGBLBBRNfog6ALzGRxmouyMac1kD1eLyQm1LpQanBsZRz/9mqiQi/KbmBY7dHDTjRTi9Eiv5G3GLXWdNKDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5206c1634f0ccff39940cd3a03c1850a72507de903e0053a12fbd07d11499e8a","last_reissued_at":"2026-07-09T00:19:19.895694Z","signature_status":"signed_v1","first_computed_at":"2026-07-09T00:19:19.895694Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"TriRoute: Unified Learned Routing for Joint Adaptive Attention, Experts, and KV-Cache Allocation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Andrii Balashov, Olena Ponomarova","submitted_at":"2026-07-07T00:12:46Z","abstract_excerpt":"Conditional computation can decouple language model quality from per-token inference cost, yet leading techniques act on a single axis in isolation: Mixture-of-Experts (MoE) sparsifies the FFN, Mixture-of-Depths (MoD) skips whole transformer blocks, and KV-cache quantization compresses attention memory. We argue these three decisions (attention resolution, expert selection, and cache bit-width) are strongly coupled and should be made jointly: a token rare enough to warrant full attention may also need high-precision caching regardless of which expert processes it. We introduce TriRoute, a sing"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.06601","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/2607.06601/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":"2607.06601","created_at":"2026-07-09T00:19:19.895763+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.06601v1","created_at":"2026-07-09T00:19:19.895763+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.06601","created_at":"2026-07-09T00:19:19.895763+00:00"},{"alias_kind":"pith_short_12","alias_value":"KIDMCY2PBTH7","created_at":"2026-07-09T00:19:19.895763+00:00"},{"alias_kind":"pith_short_16","alias_value":"KIDMCY2PBTH7HGKA","created_at":"2026-07-09T00:19:19.895763+00:00"},{"alias_kind":"pith_short_8","alias_value":"KIDMCY2P","created_at":"2026-07-09T00:19:19.895763+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/KIDMCY2PBTH7HGKAZU5AHQMFBJ","json":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ.json","graph_json":"https://pith.science/api/pith-number/KIDMCY2PBTH7HGKAZU5AHQMFBJ/graph.json","events_json":"https://pith.science/api/pith-number/KIDMCY2PBTH7HGKAZU5AHQMFBJ/events.json","paper":"https://pith.science/paper/KIDMCY2P"},"agent_actions":{"view_html":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ","download_json":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ.json","view_paper":"https://pith.science/paper/KIDMCY2P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.06601&json=true","fetch_graph":"https://pith.science/api/pith-number/KIDMCY2PBTH7HGKAZU5AHQMFBJ/graph.json","fetch_events":"https://pith.science/api/pith-number/KIDMCY2PBTH7HGKAZU5AHQMFBJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ/action/storage_attestation","attest_author":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ/action/author_attestation","sign_citation":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ/action/citation_signature","submit_replication":"https://pith.science/pith/KIDMCY2PBTH7HGKAZU5AHQMFBJ/action/replication_record"}},"created_at":"2026-07-09T00:19:19.895763+00:00","updated_at":"2026-07-09T00:19:19.895763+00:00"}