{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:RGCV7AB6DNP3HPOEX3I2HOWISI","short_pith_number":"pith:RGCV7AB6","schema_version":"1.0","canonical_sha256":"89855f803e1b5fb3bdc4bed1a3bac8923744fcd402c085a3165d3a5354309920","source":{"kind":"arxiv","id":"2605.00180","version":2},"attestation_state":"computed","paper":{"title":"RouteProfile: Graph-Based Profiling for Cold-Start LLM Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Structured profiles for LLM capabilities outperform flat ones in routing tasks and improve generalization to new models.","cross_cats":["cs.CL"],"primary_cat":"cs.NI","authors_text":"Ge Liu, Haozhen Zhang, Hongji Pu, Jiaxuan You, Jingjun Xu, Tao Feng","submitted_at":"2026-04-30T19:56:08Z","abstract_excerpt":"LLM routing is increasingly important for selecting suitable models under diverse user needs and deployment constraints, but its practical effectiveness depends on continual adaptation to emerging queries and newly released models. New-LLM integration is particularly challenging, as newly released models lack the query-response-reward interactions required for router training and cannot be profiled as directly as new queries via semantic embeddings. Existing profiles are limited: LLM-generated descriptions are often coarse, while interaction-based embeddings are costly to construct. To address"},"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":"2605.00180","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.NI","submitted_at":"2026-04-30T19:56:08Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"01b816689feb814cdf17cffa124a251f1e30ebe0575aa61a858c3cb739061b7a","abstract_canon_sha256":"c56737b6992ab6f4706ea741a2fa7afddb7c36a4ea7e427f381eb5430e521200"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:41.193264Z","signature_b64":"+n7D09iWenSqae3zSXqr1Cs/25wvxOyAEslc2OgBFmkEYeaxGrer43sr+EC9OAqTz4qlHJnYatMO+XytDYwsDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"89855f803e1b5fb3bdc4bed1a3bac8923744fcd402c085a3165d3a5354309920","last_reissued_at":"2026-05-28T01:04:41.192497Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:41.192497Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RouteProfile: Graph-Based Profiling for Cold-Start LLM Routing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Structured profiles for LLM capabilities outperform flat ones in routing tasks and improve generalization to new models.","cross_cats":["cs.CL"],"primary_cat":"cs.NI","authors_text":"Ge Liu, Haozhen Zhang, Hongji Pu, Jiaxuan You, Jingjun Xu, Tao Feng","submitted_at":"2026-04-30T19:56:08Z","abstract_excerpt":"LLM routing is increasingly important for selecting suitable models under diverse user needs and deployment constraints, but its practical effectiveness depends on continual adaptation to emerging queries and newly released models. New-LLM integration is particularly challenging, as newly released models lack the query-response-reward interactions required for router training and cannot be profiled as directly as new queries via semantic embeddings. Existing profiles are limited: LLM-generated descriptions are often coarse, while interaction-based embeddings are costly to construct. To address"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Through systematic evaluation across three representative routers under both standard and new-LLM generalization settings, we show that: (1) structured profiles consistently outperform flat ones; (2) query-level signals are more reliable than coarse domain-level signals; and (3) generalization to newly introduced models benefits most from structured profiles under trainable configurations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That evaluations on three routers and the chosen standard plus generalization settings are representative enough to elucidate the full design space and apply to other routing systems and LLMs.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"RouteProfile organizes LLM profile design into organizational form, representation type, aggregation depth, and learning configuration, with evaluations showing structured profiles outperform flat ones and aid generalization to new models.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Structured profiles for LLM capabilities outperform flat ones in routing tasks and improve generalization to new models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"7e4a5821550c211235e544a101f2e73bad52355ef793940c6a38716d43da448b"},"source":{"id":"2605.00180","kind":"arxiv","version":2},"verdict":{"id":"e246cc37-722e-462b-b387-94ba2e4704bf","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T20:09:42.748911Z","strongest_claim":"Through systematic evaluation across three representative routers under both standard and new-LLM generalization settings, we show that: (1) structured profiles consistently outperform flat ones; (2) query-level signals are more reliable than coarse domain-level signals; and (3) generalization to newly introduced models benefits most from structured profiles under trainable configurations.","one_line_summary":"RouteProfile organizes LLM profile design into organizational form, representation type, aggregation depth, and learning configuration, with evaluations showing structured profiles outperform flat ones and aid generalization to new models.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That evaluations on three routers and the chosen standard plus generalization settings are representative enough to elucidate the full design space and apply to other routing systems and LLMs.","pith_extraction_headline":"Structured profiles for LLM capabilities outperform flat ones in routing tasks and improve generalization to new models."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.00180/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T20:36:58.556216Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T18:25:13.927751Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"1ec3eead39ece01ddc4c031c7e60a94c0c1bdc2b070bfdf71aa8536757a19933"},"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":"2605.00180","created_at":"2026-05-28T01:04:41.192567+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.00180v2","created_at":"2026-05-28T01:04:41.192567+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.00180","created_at":"2026-05-28T01:04:41.192567+00:00"},{"alias_kind":"pith_short_12","alias_value":"RGCV7AB6DNP3","created_at":"2026-05-28T01:04:41.192567+00:00"},{"alias_kind":"pith_short_16","alias_value":"RGCV7AB6DNP3HPOE","created_at":"2026-05-28T01:04:41.192567+00:00"},{"alias_kind":"pith_short_8","alias_value":"RGCV7AB6","created_at":"2026-05-28T01:04:41.192567+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/RGCV7AB6DNP3HPOEX3I2HOWISI","json":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI.json","graph_json":"https://pith.science/api/pith-number/RGCV7AB6DNP3HPOEX3I2HOWISI/graph.json","events_json":"https://pith.science/api/pith-number/RGCV7AB6DNP3HPOEX3I2HOWISI/events.json","paper":"https://pith.science/paper/RGCV7AB6"},"agent_actions":{"view_html":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI","download_json":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI.json","view_paper":"https://pith.science/paper/RGCV7AB6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.00180&json=true","fetch_graph":"https://pith.science/api/pith-number/RGCV7AB6DNP3HPOEX3I2HOWISI/graph.json","fetch_events":"https://pith.science/api/pith-number/RGCV7AB6DNP3HPOEX3I2HOWISI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI/action/storage_attestation","attest_author":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI/action/author_attestation","sign_citation":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI/action/citation_signature","submit_replication":"https://pith.science/pith/RGCV7AB6DNP3HPOEX3I2HOWISI/action/replication_record"}},"created_at":"2026-05-28T01:04:41.192567+00:00","updated_at":"2026-05-28T01:04:41.192567+00:00"}