{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:SDVHQ2G6MXSIA4E3HDIII5BS4K","short_pith_number":"pith:SDVHQ2G6","schema_version":"1.0","canonical_sha256":"90ea7868de65e480709b38d0847432e28a30cb60d9d5fc937a0c42c36c16cace","source":{"kind":"arxiv","id":"2605.17106","version":1},"attestation_state":"computed","paper":{"title":"HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aashna Garg, Federico Brancasi, Jinu Jang, Shengyu Fu, Siddharth Singha Roy","submitted_at":"2026-05-16T18:19:30Z","abstract_excerpt":"Production LLM deployments increasingly maintain heterogeneous model pools spanning order-of-magnitude cost differences. Existing routers make binary strong-vs-weak decisions and couple learned parameters to specific model identities, requiring retraining whenever the catalog changes. We present HyDRA (Hybrid Dynamic Routing Architecture), a framework that predicts fine-grained, multi-dimensional capability requirements per query and matches them against configuration-defined model profiles via shortfall matching. A ModernBERT encoder with K=4 independent sigmoid heads scores each query along "},"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":"2605.17106","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-05-16T18:19:30Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"6fd979813ea013e4c9bd8d599a7d987f2e18a2ed93e07db22aa33144396921e2","abstract_canon_sha256":"1a3d880d536ff1197c6f3da0311e22abe6e71071fb93b867acba9265be6ca548"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:03:40.409675Z","signature_b64":"SVZdlql5PDS7PjjM16C6envl18QvNRu/ceFCbPN3WgZswbUHsJwssqQEE8rEE9QUO990GTeoiyqOrmbchK2BBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"90ea7868de65e480709b38d0847432e28a30cb60d9d5fc937a0c42c36c16cace","last_reissued_at":"2026-05-20T00:03:40.408838Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:03:40.408838Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HyDRA: Hybrid Dynamic Routing Architecture for Heterogeneous LLM Pools","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Aashna Garg, Federico Brancasi, Jinu Jang, Shengyu Fu, Siddharth Singha Roy","submitted_at":"2026-05-16T18:19:30Z","abstract_excerpt":"Production LLM deployments increasingly maintain heterogeneous model pools spanning order-of-magnitude cost differences. Existing routers make binary strong-vs-weak decisions and couple learned parameters to specific model identities, requiring retraining whenever the catalog changes. We present HyDRA (Hybrid Dynamic Routing Architecture), a framework that predicts fine-grained, multi-dimensional capability requirements per query and matches them against configuration-defined model profiles via shortfall matching. A ModernBERT encoder with K=4 independent sigmoid heads scores each query along "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.17106","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/2605.17106/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T22:33:23.792476Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T22:21:57.721536Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"3f55007d00332686fa5c2415661cc0ab90df5c965dc992b6744bdc1ae48eb76f"},"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":"2605.17106","created_at":"2026-05-20T00:03:40.408971+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.17106v1","created_at":"2026-05-20T00:03:40.408971+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.17106","created_at":"2026-05-20T00:03:40.408971+00:00"},{"alias_kind":"pith_short_12","alias_value":"SDVHQ2G6MXSI","created_at":"2026-05-20T00:03:40.408971+00:00"},{"alias_kind":"pith_short_16","alias_value":"SDVHQ2G6MXSIA4E3","created_at":"2026-05-20T00:03:40.408971+00:00"},{"alias_kind":"pith_short_8","alias_value":"SDVHQ2G6","created_at":"2026-05-20T00:03:40.408971+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/SDVHQ2G6MXSIA4E3HDIII5BS4K","json":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K.json","graph_json":"https://pith.science/api/pith-number/SDVHQ2G6MXSIA4E3HDIII5BS4K/graph.json","events_json":"https://pith.science/api/pith-number/SDVHQ2G6MXSIA4E3HDIII5BS4K/events.json","paper":"https://pith.science/paper/SDVHQ2G6"},"agent_actions":{"view_html":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K","download_json":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K.json","view_paper":"https://pith.science/paper/SDVHQ2G6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.17106&json=true","fetch_graph":"https://pith.science/api/pith-number/SDVHQ2G6MXSIA4E3HDIII5BS4K/graph.json","fetch_events":"https://pith.science/api/pith-number/SDVHQ2G6MXSIA4E3HDIII5BS4K/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K/action/storage_attestation","attest_author":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K/action/author_attestation","sign_citation":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K/action/citation_signature","submit_replication":"https://pith.science/pith/SDVHQ2G6MXSIA4E3HDIII5BS4K/action/replication_record"}},"created_at":"2026-05-20T00:03:40.408971+00:00","updated_at":"2026-05-20T00:03:40.408971+00:00"}