{"id":"W7117310638","doi":"10.2139/ssrn.5970054","title":"Estimating Consumer Preferences for LLMs: Evidence from LMArena","year":2025,"lang":"","type":"preprint","venue":"SSRN Electronic Journal","topic":"Text Readability and Simplification","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Counterfactual thinking; Context (archaeology); Quality (philosophy); Consumer welfare; Economic surplus; Consumer choice; Taste; Product (mathematics); Pairwise comparison","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.007809912,0.0008672189,0.00109393,0.0003496859,0.001351157,0.001676048,0.005074825,0.0007101943,0.00006181865],"category_scores_gemma":[0.002832936,0.0008284568,0.0007289157,0.0005300451,0.0002929135,0.001420841,0.001196836,0.007383357,0.00004801298],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003115218,"about_ca_system_score_gemma":0.03167995,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001439421,"about_ca_topic_score_gemma":0.002703031,"domain_scores_codex":[0.9896362,0.0008317585,0.001994557,0.002047108,0.0009703151,0.00452008],"domain_scores_gemma":[0.991279,0.003564108,0.001858597,0.001809275,0.001200487,0.0002884992],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002403285,0.0002688424,0.003614624,0.0003916723,0.001007122,0.000001649253,0.00322319,0.005939569,0.0002240339,0.05973984,0.0002024339,0.9251467],"study_design_scores_gemma":[0.0005632272,0.0004378446,0.001488165,0.002769346,0.0003528566,0.00007383393,0.0006385758,0.1913166,0.000284566,0.8004932,0.0007963044,0.0007854921],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0477414,0.02785785,0.9163952,0.00262111,0.003637254,0.0013801,0.000105862,0.000117699,0.0001434866],"genre_scores_gemma":[0.880587,0.01639491,0.09954444,0.0001508192,0.001252611,0.0002612159,0.00005521991,0.00003297128,0.001720762],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9243612,"threshold_uncertainty_score":0.9999489,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04872354047087291,"score_gpt":0.3180276697528828,"score_spread":0.2693041292820099,"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."}}