{"id":"W4409601530","doi":"10.1007/s00521-025-11164-y","title":"Self-evaluation of LLMs on challenging LLM-generated STEM MCQs","year":2025,"lang":"en","type":"article","venue":"Neural Computing and Applications","topic":"Advanced Data Storage Technologies","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Computational Science and Engineering; Computer science; Mathematics education; Psychology; Machine learning","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":[],"consensus_categories":[],"category_scores_codex":[0.0002510492,0.00009028115,0.0001105513,0.0001290119,0.000219129,0.0000393494,0.0003961862,0.00004188235,2.459773e-7],"category_scores_gemma":[0.00002046815,0.00008683679,0.00001838343,0.0005512979,0.00003712614,0.00009996772,0.0002472427,0.0001153241,0.00000245613],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003523443,"about_ca_system_score_gemma":0.00002545905,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001944405,"about_ca_topic_score_gemma":6.965178e-7,"domain_scores_codex":[0.9991735,0.00004427721,0.0001774303,0.0003133987,0.0001674726,0.0001239703],"domain_scores_gemma":[0.9991445,0.0001261916,0.00009474398,0.0004830467,0.0001329526,0.00001855918],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00000103837,0.00007770395,0.00006519693,0.00003331631,0.00001290028,2.593556e-7,0.0001133424,0.01831665,0.001395051,0.3167767,0.0001280783,0.6630797],"study_design_scores_gemma":[0.0002099136,0.0000354884,0.0004209919,0.00002735576,0.00001357739,0.000002290913,0.00006779569,0.984292,0.00739767,0.005672465,0.001768355,0.00009210766],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1029742,0.0005305927,0.8926497,0.001112759,0.00007969496,0.0004289831,0.000004018734,0.0007965673,0.001423544],"genre_scores_gemma":[0.9772828,0.00002108278,0.02252994,0.00007803909,0.00001786123,0.000039887,0.00000471197,0.000003579188,0.00002205497],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9659753,"threshold_uncertainty_score":0.3541102,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03904450454822871,"score_gpt":0.3134778578845829,"score_spread":0.2744333533363542,"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."}}