{"id":"W7115709189","doi":"10.48448/gb1y-9631","title":"Understanding How a Language Model Assesses the Quality of Randomized Controlled Trials: Applying Shapley Additive Explanations to Encoder Transformer Classification Models","year":2025,"lang":"","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University","funders":"","keywords":"Language model; Workflow; Quality (philosophy); Test (biology); Critical appraisal; Deep learning","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","scholarly_communication","insufficient_payload"],"consensus_categories":["metaresearch","metaepi_narrow","sts"],"category_scores_codex":[0.1030975,0.00174173,0.01223336,0.005734023,0.002122064,0.001651055,0.003363499,0.0007883073,0.00120418],"category_scores_gemma":[0.05544327,0.001121512,0.002777988,0.006642227,0.006141079,0.002155178,0.0002506197,0.001273056,0.00008281064],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003042274,"about_ca_system_score_gemma":0.01171245,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006484055,"about_ca_topic_score_gemma":0.002595248,"domain_scores_codex":[0.9715976,0.01021389,0.007282446,0.002850525,0.00623081,0.001824772],"domain_scores_gemma":[0.9430961,0.04110343,0.009212979,0.002384869,0.003544949,0.000657666],"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":[0.0568831,0.0007679572,0.000001788152,0.0002715745,0.003169002,0.000001795185,0.01513789,0.1631376,0.03573425,0.7203425,0.002598685,0.001953824],"study_design_scores_gemma":[0.144831,0.00005114792,0.000001634949,0.001034045,0.002826079,0.000001506859,0.05460439,0.7679046,0.0006625429,0.02706177,0.00008138494,0.0009399822],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0001355746,0.0009610559,0.8391069,0.004043998,0.0005338017,0.0260289,0.00551329,0.0002069914,0.1234695],"genre_scores_gemma":[0.9585086,0.0009031378,0.0156199,0.0004223469,0.0002791204,0.007251916,0.0004361552,0.0003622504,0.01621659],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.958373,"threshold_uncertainty_score":0.9997088,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3430963393206387,"score_gpt":0.4309888300504897,"score_spread":0.087892490729851,"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."}}