{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005940073,0.0004794933,0.0006881308,0.001087629,0.0006688409,0.00162083,0.001272412,0.001095928,0.02137853],"category_scores_gemma":[0.03286634,0.000367758,0.001020199,0.001733974,0.0006182053,0.002208312,0.000977729,0.00117767,0.003568731],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001313432,"about_ca_system_score_gemma":0.0004672935,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03354764,"about_ca_topic_score_gemma":0.02639672,"domain_scores_codex":[0.9970034,0.001994926,0.00009948016,0.0003039611,0.0004732311,0.0001250075],"domain_scores_gemma":[0.9300011,0.05979674,0.002927514,0.004625239,0.002084998,0.0005643533],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.02354033,0.004614191,0.6142395,0.001356343,0.00167752,0.0004874461,0.004246432,0.03213683,0.005916397,0.01356915,0.02177596,0.27644],"study_design_scores_gemma":[0.001251836,0.002683904,0.8382248,0.0003010296,0.001238634,0.000183863,0.006674929,0.09779128,0.008494916,0.01611423,0.02681825,0.0002222546],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9812984,0.0005036718,0.00423782,0.0005151525,0.00001370259,0.00007254531,0.001771491,0.0001620477,0.0114252],"genre_scores_gemma":[0.9885352,0.0001968715,0.003578202,0.0001715041,0.00001387729,0.00005549083,0.002751744,0.00007138059,0.004625659],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.03354764,"threshold_uncertainty_score":0.07151824,"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."}}