{"id":"W4417041459","doi":"10.48550/arxiv.2504.20006","title":"Chatbot Arena Meets Nuggets: Towards Explanations and Diagnostics in the Evaluation of LLM Responses","year":2025,"lang":"en","type":"preprint","venue":"ArXiv.org","topic":"AI in Service Interactions","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Chatbot; Context (archaeology); Quality (philosophy); Code (set theory); Work (physics)","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001731216,0.0002009469,0.0002537327,0.0003544434,0.0001087331,0.0001029205,0.001291311,0.0001656478,0.00001889087],"category_scores_gemma":[0.001785161,0.0001697431,0.00006911191,0.0004266489,0.00006540344,0.00026876,0.001057407,0.0004761946,0.00001149846],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001596096,"about_ca_system_score_gemma":0.0005653029,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005064625,"about_ca_topic_score_gemma":0.0007546414,"domain_scores_codex":[0.9975398,0.0006598262,0.0004790122,0.000485268,0.0006545966,0.0001814803],"domain_scores_gemma":[0.9961898,0.00180807,0.0002870812,0.001132516,0.0005475109,0.00003503731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001673474,0.003267605,0.5134607,0.001637515,0.0008382613,0.0001109521,0.1372033,0.02295418,0.0006574155,0.05494905,0.01791038,0.2468433],"study_design_scores_gemma":[0.0004039651,0.00006617634,0.9479616,0.0007954171,0.0001477484,0.00001075551,0.0009308709,0.0351619,0.0007595901,0.009923927,0.003533994,0.000304079],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9674276,0.001596744,0.008732899,0.01612439,0.001345265,0.0008347408,0.0001011922,0.00007319761,0.003764007],"genre_scores_gemma":[0.9954551,0.0006976689,0.002571027,0.0007280527,0.00007143909,0.0003402811,0.00003636956,0.00000712307,0.0000928736],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4345009,"threshold_uncertainty_score":0.6921922,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1453284930399543,"score_gpt":0.3884572676797482,"score_spread":0.2431287746397939,"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."}}