{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:NWMUA2ENQUCYOF35FY5KKOQOTT","short_pith_number":"pith:NWMUA2EN","schema_version":"1.0","canonical_sha256":"6d9940688d850587177d2e3aa53a0e9cf783d32d1afdfa276114b1035a37e2f5","source":{"kind":"arxiv","id":"2502.02743","version":1},"attestation_state":"computed","paper":{"title":"LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned Dynamic Routing","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Yang Li","submitted_at":"2025-02-04T22:09:43Z","abstract_excerpt":"The rapid advancement in large language models (LLMs) has brought forth a diverse range of models with varying capabilities that excel in different tasks and domains. However, selecting the optimal LLM for user queries often involves a challenging trade-off between accuracy and cost, a problem exacerbated by the diverse demands of individual queries. In this work, we present a novel framework that formulates the LLM selection process as a multi-armed bandit problem, enabling dynamic and intelligent routing of queries to the most appropriate model. Our approach incorporates a preference-conditi"},"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":"2502.02743","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T22:09:43Z","cross_cats_sorted":[],"title_canon_sha256":"16bc6cf0e937cc200627e6c81977f3bdd2c1dcb9ee8bb4eedee0730a0fa4c6ca","abstract_canon_sha256":"3d3c80ae2481214f87069b5343dff1e9775661c219d9fa0019694838df010e6f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:09:54.568668Z","signature_b64":"bbcrn74I+KbCNm5jVqLxui/4GghjKdLxyW1Q061dSXteMBMI4UHmsDzxZ5S+PFtewnpq4PAQD36U9jZx/+FZCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d9940688d850587177d2e3aa53a0e9cf783d32d1afdfa276114b1035a37e2f5","last_reissued_at":"2026-07-05T10:09:54.568329Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:09:54.568329Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM Bandit: Cost-Efficient LLM Generation via Preference-Conditioned Dynamic Routing","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Yang Li","submitted_at":"2025-02-04T22:09:43Z","abstract_excerpt":"The rapid advancement in large language models (LLMs) has brought forth a diverse range of models with varying capabilities that excel in different tasks and domains. However, selecting the optimal LLM for user queries often involves a challenging trade-off between accuracy and cost, a problem exacerbated by the diverse demands of individual queries. In this work, we present a novel framework that formulates the LLM selection process as a multi-armed bandit problem, enabling dynamic and intelligent routing of queries to the most appropriate model. Our approach incorporates a preference-conditi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.02743","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/2502.02743/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":"2502.02743","created_at":"2026-07-05T10:09:54.568382+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.02743v1","created_at":"2026-07-05T10:09:54.568382+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.02743","created_at":"2026-07-05T10:09:54.568382+00:00"},{"alias_kind":"pith_short_12","alias_value":"NWMUA2ENQUCY","created_at":"2026-07-05T10:09:54.568382+00:00"},{"alias_kind":"pith_short_16","alias_value":"NWMUA2ENQUCYOF35","created_at":"2026-07-05T10:09:54.568382+00:00"},{"alias_kind":"pith_short_8","alias_value":"NWMUA2EN","created_at":"2026-07-05T10:09:54.568382+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2502.18036","citing_title":"Harnessing Multiple Large Language Models: A Survey on LLM Ensemble","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2505.12601","citing_title":"Rethinking Predictive Modeling for LLM Routing: When Simple kNN Beats Complex Learned Routers","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2602.07303","citing_title":"KRONE: Scalable LLM-Augmented Log Anomaly Detection via Hierarchical Abstraction","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03426","citing_title":"Replacing Parameters with Preferences: Federated Alignment of Heterogeneous Vision-Language Models","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16583","citing_title":"POLAR: Online Learning for LoRA Adapter Caching and Routing in Edge LLM Serving","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT","json":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT.json","graph_json":"https://pith.science/api/pith-number/NWMUA2ENQUCYOF35FY5KKOQOTT/graph.json","events_json":"https://pith.science/api/pith-number/NWMUA2ENQUCYOF35FY5KKOQOTT/events.json","paper":"https://pith.science/paper/NWMUA2EN"},"agent_actions":{"view_html":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT","download_json":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT.json","view_paper":"https://pith.science/paper/NWMUA2EN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.02743&json=true","fetch_graph":"https://pith.science/api/pith-number/NWMUA2ENQUCYOF35FY5KKOQOTT/graph.json","fetch_events":"https://pith.science/api/pith-number/NWMUA2ENQUCYOF35FY5KKOQOTT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT/action/storage_attestation","attest_author":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT/action/author_attestation","sign_citation":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT/action/citation_signature","submit_replication":"https://pith.science/pith/NWMUA2ENQUCYOF35FY5KKOQOTT/action/replication_record"}},"created_at":"2026-07-05T10:09:54.568382+00:00","updated_at":"2026-07-05T10:09:54.568382+00:00"}