{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GD73VJE45W5M5FWNVIC36XOOCA","short_pith_number":"pith:GD73VJE4","schema_version":"1.0","canonical_sha256":"30ffbaa49cedbace96cdaa05bf5dce102aa9e30e6888527326314ef176a849ff","source":{"kind":"arxiv","id":"2502.18482","version":1},"attestation_state":"computed","paper":{"title":"MixLLM: Dynamic Routing in Mixed Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.DB","cs.IR"],"primary_cat":"cs.CL","authors_text":"Haifeng Chen, Wei Cheng, Wenchao Yu, Xinyuan Wang, Xujiang Zhao, Yanchi Liu, Yanjie Fu, Zhengzhang Chen","submitted_at":"2025-02-09T02:26:15Z","abstract_excerpt":"Large Language Models (LLMs) exhibit potential artificial generic intelligence recently, however, their usage is costly with high response latency. Given mixed LLMs with their own strengths and weaknesses, LLM routing aims to identify the most suitable model for each query in the stream to maximize response quality and minimize cost and latency. However, the challenges involve: (1) dynamic trade-offs among quality, cost, and latency; (2) enabling continual learning in deployed systems; and (3) navigating a varying (e.g., new LLM addition or old LLM removal) set of LLM candidates over time. To "},"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":"2502.18482","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-09T02:26:15Z","cross_cats_sorted":["cs.AI","cs.DB","cs.IR"],"title_canon_sha256":"11f1f79aabc565b5db158eeef86c096f37f2c7b4d5ac09630a5cbd739da16414","abstract_canon_sha256":"a975a44082ff736b6a26561bba127c2952a6fd536b859d727fc399586b911508"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:19:58.599016Z","signature_b64":"H1MhiAGFCnisjtbH4mzrDLLDts0o1E4QsdeHVBLgblFBlp+3ClUVGsiuzztAc9i0omDy6h3sQFzm4Qo/mTSeCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"30ffbaa49cedbace96cdaa05bf5dce102aa9e30e6888527326314ef176a849ff","last_reissued_at":"2026-07-05T10:19:58.598525Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:19:58.598525Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MixLLM: Dynamic Routing in Mixed Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.DB","cs.IR"],"primary_cat":"cs.CL","authors_text":"Haifeng Chen, Wei Cheng, Wenchao Yu, Xinyuan Wang, Xujiang Zhao, Yanchi Liu, Yanjie Fu, Zhengzhang Chen","submitted_at":"2025-02-09T02:26:15Z","abstract_excerpt":"Large Language Models (LLMs) exhibit potential artificial generic intelligence recently, however, their usage is costly with high response latency. Given mixed LLMs with their own strengths and weaknesses, LLM routing aims to identify the most suitable model for each query in the stream to maximize response quality and minimize cost and latency. However, the challenges involve: (1) dynamic trade-offs among quality, cost, and latency; (2) enabling continual learning in deployed systems; and (3) navigating a varying (e.g., new LLM addition or old LLM removal) set of LLM candidates over time. To "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18482","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/2502.18482/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":"2502.18482","created_at":"2026-07-05T10:19:58.598582+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.18482v1","created_at":"2026-07-05T10:19:58.598582+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18482","created_at":"2026-07-05T10:19:58.598582+00:00"},{"alias_kind":"pith_short_12","alias_value":"GD73VJE45W5M","created_at":"2026-07-05T10:19:58.598582+00:00"},{"alias_kind":"pith_short_16","alias_value":"GD73VJE45W5M5FWN","created_at":"2026-07-05T10:19:58.598582+00:00"},{"alias_kind":"pith_short_8","alias_value":"GD73VJE4","created_at":"2026-07-05T10:19:58.598582+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.18774","citing_title":"RouteJudge: An Open Platform for Reproducible and Preference-Aware LLM Routing","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06924","citing_title":"From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routing","ref_index":126,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25424","citing_title":"SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learning","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17288","citing_title":"When Efficiency Backfires: Cascading LLMs Trigger Cascade Failure under Adversarial Attack","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2603.21354","citing_title":"The Workload-Router-Pool Architecture for LLM Inference Optimization: A Vision Paper from the vLLM Semantic Router Project","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA","json":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA.json","graph_json":"https://pith.science/api/pith-number/GD73VJE45W5M5FWNVIC36XOOCA/graph.json","events_json":"https://pith.science/api/pith-number/GD73VJE45W5M5FWNVIC36XOOCA/events.json","paper":"https://pith.science/paper/GD73VJE4"},"agent_actions":{"view_html":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA","download_json":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA.json","view_paper":"https://pith.science/paper/GD73VJE4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.18482&json=true","fetch_graph":"https://pith.science/api/pith-number/GD73VJE45W5M5FWNVIC36XOOCA/graph.json","fetch_events":"https://pith.science/api/pith-number/GD73VJE45W5M5FWNVIC36XOOCA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA/action/storage_attestation","attest_author":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA/action/author_attestation","sign_citation":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA/action/citation_signature","submit_replication":"https://pith.science/pith/GD73VJE45W5M5FWNVIC36XOOCA/action/replication_record"}},"created_at":"2026-07-05T10:19:58.598582+00:00","updated_at":"2026-07-05T10:19:58.598582+00:00"}