{"paper":{"title":"Small Language Models for Inference-Capability Routing: Effects of Model Capacity and Prompt Optimization on Downstream Quality and Efficiency","arxiv_id":"2611.16414","primary_cat":"cs.AI","submitted_at":"2026-11-12T07:25:46+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Optional reasoning, retrieval, web search, and tool use can improve large language model responses but incur unnecessary computation when activated indiscriminately.","hedge":{"p":0.59,"label":"hedged_possibility"},"own_p":0.06,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.87,"id":"s1","own":false,"role":"contribution","text":"This study investigates whether prompt optimization can compensate for limited router capacity and whether classification gains translate into downstream efficiency.","hedge":{"p":0.61,"label":"unclear"},"own_p":0.39,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.12,"method":0.0,"result":0.0,"background":0.01,"limitation":0.0,"future_work":0.0,"contribution":0.87,"interpretation":0.0}},{"i":2,"p":0.9,"id":"s2","own":true,"role":"method","text":"The proposed framework uses prompt-driven small language models to make four independent capability decisions, allowing multiple capabilities to be selected per query.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.79,"main_claim_p":0.04,"quantitative":true,"quantitative_p":0.6,"role_probabilities":{"gap":0.0,"method":0.9,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.1,"interpretation":0.0}},{"i":3,"p":0.99,"id":"s3","own":true,"role":"method","text":"Qwen2.5 routers ranging from 1.5B to 14B parameters are compared across five prompting strategies, followed by an evaluation of capability-specific optimized configurations using a fixed GPT-OSS-120B downstream model on 127 questions.","hedge":{"p":0.86,"label":"asserted"},"own_p":0.95,"main_claim_p":0.04,"quantitative":true,"quantitative_p":0.98,"role_probabilities":{"gap":0.0,"method":0.99,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.01,"interpretation":0.0}},{"i":4,"p":1.0,"id":"s4","own":true,"role":"result","text":"Under minimal prompting, increasing router capacity does not consistently improve classification performance.","hedge":{"p":0.95,"label":"asserted"},"own_p":0.92,"main_claim_p":0.08,"quantitative":false,"quantitative_p":0.23,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":5,"p":0.96,"id":"s5","own":true,"role":"result","text":"Prompt optimization narrows the capacity gap for reasoning and calculator routing, whereas RAG and web search retain larger benefits from higher capacity.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.93,"main_claim_p":0.15,"quantitative":true,"quantitative_p":0.88,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.96,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.04}},{"i":6,"p":0.93,"id":"s6","own":true,"role":"interpretation","text":"Similar classification F1 scores do not necessarily imply equivalent downstream behavior: false-positive activations increase unnecessary computation, while missed capabilities can reduce answer quality.","hedge":{"p":0.77,"label":"asserted"},"own_p":0.67,"main_claim_p":0.16,"quantitative":false,"quantitative_p":0.16,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.07,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.93}},{"i":7,"p":1.0,"id":"s7","own":true,"role":"result","text":"The 14B router achieves a mean BERTScore F1 of 0.8408, compared with 0.8418 when all capabilities are enabled, while reducing average downstream token consumption by 68.7\\% (8,837 to 2,764 tokens) and median downstream latency by 80.5\\% (9.79 to 1.91 seconds).","hedge":{"p":1.0,"label":"asserted"},"own_p":0.96,"main_claim_p":0.2,"quantitative":true,"quantitative_p":0.99,"role_probabilities":{"gap":0.0,"method":0.0,"result":1.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":8,"p":0.94,"id":"s8","own":true,"role":"limitation","text":"These savings exclude router overhead, which is measured separately in a different inference environment and increases with router capacity.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.92,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.08,"role_probabilities":{"gap":0.0,"method":0.04,"result":0.02,"background":0.0,"limitation":0.94,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":9,"p":0.84,"id":"s9","own":true,"role":"interpretation","text":"The findings demonstrate that the relationship between router capacity, prompt design, and downstream efficiency is capability-dependent.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.91,"main_claim_p":0.53,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.16,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.84}},{"i":10,"p":0.95,"id":"s10","own":false,"role":"interpretation","text":"Consequently, router selection must consider error profiles and routing overhead alongside classification performance.","hedge":{"p":0.96,"label":"asserted"},"own_p":0.48,"main_claim_p":0.48,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.05,"contribution":0.0,"interpretation":0.95}}]},"compiled":{"main_claim":{"p":0.72,"text":"The findings demonstrate that the relationship between router capacity, prompt design, and downstream efficiency is capability-dependent.","location":"abstract s9","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":8316,"cost_usd_reported":0.0},"compiled_at":"2026-11-13","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.03,"states_limitation":0.94,"human_participants":0.04,"controlled_comparison":0.91,"randomized_assignment":0.02,"states_quantitative_result":0.99,"compares_with_prior_results":0.89,"mentions_code_or_data_availability":0.02},"evidential_basis":{"p":0.98,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.98,"simulation_or_computation":0.02}},"contribution_type":{"p":0.99,"value":"empirical study","probabilities":{"review":0.0,"method_or_tool":0.01,"empirical_study":0.99,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":0.8,"text":"The proposed framework uses prompt-driven small language models to make four independent capability decisions, allowing multiple capabilities to be selected per query.","location":"abstract s2"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":11,"sentences_in_abstract":11},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}