{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:THN7ZF47F6DMPDVLFTJTGUS2ZC","short_pith_number":"pith:THN7ZF47","schema_version":"1.0","canonical_sha256":"99dbfc979f2f86c78eab2cd333525ac8bab27be90eed77637c6705ba22515ccc","source":{"kind":"arxiv","id":"2505.19435","version":1},"attestation_state":"computed","paper":{"title":"Route to Reason: Adaptive Routing for LLM and Reasoning Strategy Selection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kai Zhang, Yupeng Han, Yuze Zhao, Zhihong Pan","submitted_at":"2025-05-26T02:53:17Z","abstract_excerpt":"The inherent capabilities of a language model (LM) and the reasoning strategies it employs jointly determine its performance in reasoning tasks. While test-time scaling is regarded as an effective approach to tackling complex reasoning tasks, it incurs substantial computational costs and often leads to \"overthinking\", where models become trapped in \"thought pitfalls\". To address this challenge, we propose Route-To-Reason (RTR), a novel unified routing framework that dynamically allocates both LMs and reasoning strategies according to task difficulty under budget constraints. RTR learns compres"},"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":"2505.19435","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-05-26T02:53:17Z","cross_cats_sorted":[],"title_canon_sha256":"5cc7b816ad9727c5a01a4ad8290f2851ad989265910cc980d9334f7372a1c24c","abstract_canon_sha256":"91a4a558bbd4e04690325230458972d25a8f36dddd8ad638d098f1f5d2c5b1f4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:30.184595Z","signature_b64":"s8m5zlm1r4yesGkBaPW+mQAvpeEjhErdT8oTBaT8ZSu0hAeRMErt2hkCrN4kN8rvOjnPU5RmoJud6noKcoY0AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"99dbfc979f2f86c78eab2cd333525ac8bab27be90eed77637c6705ba22515ccc","last_reissued_at":"2026-07-05T11:09:30.184090Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:30.184090Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Route to Reason: Adaptive Routing for LLM and Reasoning Strategy Selection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Kai Zhang, Yupeng Han, Yuze Zhao, Zhihong Pan","submitted_at":"2025-05-26T02:53:17Z","abstract_excerpt":"The inherent capabilities of a language model (LM) and the reasoning strategies it employs jointly determine its performance in reasoning tasks. While test-time scaling is regarded as an effective approach to tackling complex reasoning tasks, it incurs substantial computational costs and often leads to \"overthinking\", where models become trapped in \"thought pitfalls\". To address this challenge, we propose Route-To-Reason (RTR), a novel unified routing framework that dynamically allocates both LMs and reasoning strategies according to task difficulty under budget constraints. RTR learns compres"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.19435","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/2505.19435/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":"2505.19435","created_at":"2026-07-05T11:09:30.184150+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.19435v1","created_at":"2026-07-05T11:09:30.184150+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.19435","created_at":"2026-07-05T11:09:30.184150+00:00"},{"alias_kind":"pith_short_12","alias_value":"THN7ZF47F6DM","created_at":"2026-07-05T11:09:30.184150+00:00"},{"alias_kind":"pith_short_16","alias_value":"THN7ZF47F6DMPDVL","created_at":"2026-07-05T11:09:30.184150+00:00"},{"alias_kind":"pith_short_8","alias_value":"THN7ZF47","created_at":"2026-07-05T11:09:30.184150+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"internal_anchor_count":10,"sample":[{"citing_arxiv_id":"2606.23181","citing_title":"DART: Draft-Agreement Routing for Training-Free Adaptive Thinking Budgets in Hybrid Reasoning Models","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31285","citing_title":"Spatial Reasoning via Modality Switching Between Language and Symbolic Representation","ref_index":77,"is_internal_anchor":true},{"citing_arxiv_id":"2607.00053","citing_title":"SWE-Router: Routing in Multi-turn Agentic Software Engineering Tasks","ref_index":69,"is_internal_anchor":true},{"citing_arxiv_id":"2606.06924","citing_title":"From Sampled Outcomes to Capability Distributions: Rethinking Supervision for LLM Routing","ref_index":95,"is_internal_anchor":true},{"citing_arxiv_id":"2606.04378","citing_title":"DLLG: Dynamic Logit-Level Gating of LLM Experts","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31285","citing_title":"Spatial Reasoning via Modality Switching Between Language and Symbolic Representation","ref_index":77,"is_internal_anchor":true},{"citing_arxiv_id":"2605.25424","citing_title":"SeqRoute: Global Budget-Aware Sequential LLM Routing via Offline Reinforcement Learning","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06110","citing_title":"On Time, Within Budget: Constraint-Driven Online Resource Allocation for Agentic Workflows","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2605.06110","citing_title":"On Time, Within Budget: Constraint-Driven Online Resource Allocation for Agentic Workflows","ref_index":10,"is_internal_anchor":true},{"citing_arxiv_id":"2604.14853","citing_title":"Adaptive Test-Time Compute Allocation for Reasoning LLMs via Constrained Policy Optimization","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC","json":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC.json","graph_json":"https://pith.science/api/pith-number/THN7ZF47F6DMPDVLFTJTGUS2ZC/graph.json","events_json":"https://pith.science/api/pith-number/THN7ZF47F6DMPDVLFTJTGUS2ZC/events.json","paper":"https://pith.science/paper/THN7ZF47"},"agent_actions":{"view_html":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC","download_json":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC.json","view_paper":"https://pith.science/paper/THN7ZF47","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.19435&json=true","fetch_graph":"https://pith.science/api/pith-number/THN7ZF47F6DMPDVLFTJTGUS2ZC/graph.json","fetch_events":"https://pith.science/api/pith-number/THN7ZF47F6DMPDVLFTJTGUS2ZC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC/action/storage_attestation","attest_author":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC/action/author_attestation","sign_citation":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC/action/citation_signature","submit_replication":"https://pith.science/pith/THN7ZF47F6DMPDVLFTJTGUS2ZC/action/replication_record"}},"created_at":"2026-07-05T11:09:30.184150+00:00","updated_at":"2026-07-05T11:09:30.184150+00:00"}