{"id":"W4407793840","doi":"10.1098/rsos.241313","title":"Text understanding in GPT-4 versus humans","year":2025,"lang":"en","type":"article","venue":"Royal Society Open Science","topic":"Text Readability and Simplification","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute; McGill University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Biology; Computer science; Computational biology","routes":{"ca_aff":true,"ca_fund":true,"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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.001853648,0.00009304033,0.0001256205,0.00006285545,0.0006632937,0.001048769,0.004431127,0.00005243936,0.00001953752],"category_scores_gemma":[0.0001274905,0.00008948468,0.00005919841,0.002020803,0.0004317321,0.00105434,0.001522455,0.00015522,0.00002640952],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008016282,"about_ca_system_score_gemma":0.00056491,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003788446,"about_ca_topic_score_gemma":0.0002082812,"domain_scores_codex":[0.9983948,0.00003695963,0.0001999456,0.0006467314,0.0003594866,0.0003620941],"domain_scores_gemma":[0.9989686,0.0001709996,0.00004877726,0.0006858686,0.00005687493,0.00006890595],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001102081,0.00007676239,0.003216755,0.000009890531,0.000004554191,7.107166e-7,0.002941317,0.0005448775,0.0004009579,0.9806755,0.002110261,0.01000737],"study_design_scores_gemma":[0.003472131,0.0001960225,0.2171197,0.0001664398,0.00001207794,0.000001144798,0.009016384,0.608738,0.003137823,0.1392208,0.01801118,0.0009082771],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04712728,0.00005263941,0.8118092,0.006447358,0.000855453,0.0005647476,0.00000161944,0.0001133674,0.1330283],"genre_scores_gemma":[0.9853408,0.000005432981,0.0130201,0.0004205285,0.000008873777,0.00001630189,5.843347e-7,0.000002001249,0.001185326],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9382136,"threshold_uncertainty_score":0.9999883,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06907462797592029,"score_gpt":0.3414232356844297,"score_spread":0.2723486077085094,"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."}}