{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:FVQIWQ7SCCKC3UGBX4BJIZRIOF","short_pith_number":"pith:FVQIWQ7S","schema_version":"1.0","canonical_sha256":"2d608b43f210942dd0c1bf029466287178369f8b6990f9145438812093e85b69","source":{"kind":"arxiv","id":"2508.21141","version":2},"attestation_state":"computed","paper":{"title":"Adaptive LLM Routing under Budget Constraints","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chaitanya Devaguptapu, Pranoy Panda, Raghav Magazine, Sho Takemori, Vishal Sharma","submitted_at":"2025-08-28T18:18:19Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized natural language processing, but their varying capabilities and costs pose challenges in practical applications. LLM routing addresses this by dynamically selecting the most suitable LLM for each query/task. Previous approaches treat this as a supervised learning problem, assuming complete knowledge of optimal query-LLM pairings. However, real-world scenarios lack such comprehensive mappings and face evolving user queries. We thus propose to study LLM routing as a contextual bandit problem, enabling adaptive decision-making using bandit feedback"},"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":"2508.21141","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-28T18:18:19Z","cross_cats_sorted":[],"title_canon_sha256":"0811a5dc4a11b6ef3792ceb8d1dca3b731a75ed6610779e57a7ca5e37f068da6","abstract_canon_sha256":"1f242023fc1c9f88d8efb5e1b4c2beb5f06eb1054eda29e9fb60a57a953a6738"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:11.091446Z","signature_b64":"ZI+GGu5cBzKLo2YGcowuvcLyb71yAzP75kwC3gOmdvKQFugyCBUX+/7rzlSVx+LVADoo7AgbADkn/KkJpJsNAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d608b43f210942dd0c1bf029466287178369f8b6990f9145438812093e85b69","last_reissued_at":"2026-07-05T12:07:11.091033Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:11.091033Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adaptive LLM Routing under Budget Constraints","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chaitanya Devaguptapu, Pranoy Panda, Raghav Magazine, Sho Takemori, Vishal Sharma","submitted_at":"2025-08-28T18:18:19Z","abstract_excerpt":"Large Language Models (LLMs) have revolutionized natural language processing, but their varying capabilities and costs pose challenges in practical applications. LLM routing addresses this by dynamically selecting the most suitable LLM for each query/task. Previous approaches treat this as a supervised learning problem, assuming complete knowledge of optimal query-LLM pairings. However, real-world scenarios lack such comprehensive mappings and face evolving user queries. We thus propose to study LLM routing as a contextual bandit problem, enabling adaptive decision-making using bandit feedback"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.21141","kind":"arxiv","version":2},"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/2508.21141/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":"2508.21141","created_at":"2026-07-05T12:07:11.091086+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.21141v2","created_at":"2026-07-05T12:07:11.091086+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.21141","created_at":"2026-07-05T12:07:11.091086+00:00"},{"alias_kind":"pith_short_12","alias_value":"FVQIWQ7SCCKC","created_at":"2026-07-05T12:07:11.091086+00:00"},{"alias_kind":"pith_short_16","alias_value":"FVQIWQ7SCCKC3UGB","created_at":"2026-07-05T12:07:11.091086+00:00"},{"alias_kind":"pith_short_8","alias_value":"FVQIWQ7S","created_at":"2026-07-05T12:07:11.091086+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.17949","citing_title":"RouteBalance: Fused Model Routing and Load Balancing for Heterogeneous LLM Serving","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27999","citing_title":"Learning to Assign Prediction Tasks to Agents with Capacity Constraints","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.00136","citing_title":"ParetoBandit: Budget-Paced Adaptive Routing for Non-Stationary LLM Serving","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06110","citing_title":"On Time, Within Budget: Constraint-Driven Online Resource Allocation for Agentic Workflows","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05657","citing_title":"Retrieval-Conditioned Topology Selection with Provable Budget Conservation for Multi-Agent Code Generation","ref_index":77,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02241","citing_title":"Zero-Shot Confidence Estimation for Small LLMs: When Supervised Baselines Aren't Worth Training","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07180","citing_title":"Learning Agent Routing From Early Experience","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06110","citing_title":"On Time, Within Budget: Constraint-Driven Online Resource Allocation for Agentic Workflows","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14853","citing_title":"Adaptive Test-Time Compute Allocation for Reasoning LLMs via Constrained Policy Optimization","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF","json":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF.json","graph_json":"https://pith.science/api/pith-number/FVQIWQ7SCCKC3UGBX4BJIZRIOF/graph.json","events_json":"https://pith.science/api/pith-number/FVQIWQ7SCCKC3UGBX4BJIZRIOF/events.json","paper":"https://pith.science/paper/FVQIWQ7S"},"agent_actions":{"view_html":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF","download_json":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF.json","view_paper":"https://pith.science/paper/FVQIWQ7S","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.21141&json=true","fetch_graph":"https://pith.science/api/pith-number/FVQIWQ7SCCKC3UGBX4BJIZRIOF/graph.json","fetch_events":"https://pith.science/api/pith-number/FVQIWQ7SCCKC3UGBX4BJIZRIOF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF/action/storage_attestation","attest_author":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF/action/author_attestation","sign_citation":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF/action/citation_signature","submit_replication":"https://pith.science/pith/FVQIWQ7SCCKC3UGBX4BJIZRIOF/action/replication_record"}},"created_at":"2026-07-05T12:07:11.091086+00:00","updated_at":"2026-07-05T12:07:11.091086+00:00"}