{"id":"W4417509732","doi":"10.1109/aibthings66987.2025.11296175","title":"From Large to Small Language Models for Balanced Performance and Benchmarking","year":2025,"lang":"","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Lakehead University","funders":"","keywords":"Benchmarking; Inference; Baseline (sea); Sustainable development; Language model; Natural language understanding; Natural language; Comprehension","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003817206,0.001483475,0.0008733206,0.001039512,0.0007062619,0.00237975,0.002748495,0.001420751,0.005675667],"category_scores_gemma":[0.01884511,0.0005931194,0.001019002,0.001408802,0.0008063106,0.004861919,0.002058816,0.002515896,0.00308263],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001607346,"about_ca_system_score_gemma":0.002601127,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01324053,"about_ca_topic_score_gemma":0.01814326,"domain_scores_codex":[0.9966123,0.00130515,0.0002604959,0.000725179,0.0008213421,0.0002755395],"domain_scores_gemma":[0.9946063,0.002613925,0.0001505882,0.001511449,0.0008418277,0.00027596],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0019677,0.00116294,0.01190781,0.002034087,0.0007604288,0.000466365,0.0009587923,0.5122013,0.02166474,0.03272664,0.09510434,0.3190448],"study_design_scores_gemma":[0.0001698647,0.0002459663,0.001011553,0.00008444343,0.00007503452,0.00007474406,0.0003090504,0.9440835,0.01435307,0.02014193,0.01940825,0.00004247135],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4454836,0.005993187,0.3936178,0.005519398,0.001610021,0.0009712506,0.01146313,0.08481067,0.05053098],"genre_scores_gemma":[0.705227,0.0009507161,0.26984,0.0008513607,0.00009846492,0.0007468052,0.01355261,0.004282507,0.004450453],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01324053,"threshold_uncertainty_score":0.02632689,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01926470067081895,"score_gpt":0.2546951607917388,"score_spread":0.2354304601209198,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}