{"id":"W4405967087","doi":"10.1101/2024.12.26.24319521","title":"Citation Contamination of Systematic Review Literature in the Life Sciences by Paper Mills","year":2024,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Academic Writing and Publishing","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Calgary","funders":"","keywords":"Contamination; Citation; Environmental science; Library science; Computer science; Biology; Ecology","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[{"model":"gemma","categories":["metaresearch","research_integrity","bibliometrics"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low","status":"direct model label, unvalidated"},{"model":"gpt","categories":["metaresearch","bibliometrics","research_integrity"],"domain":"methods","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low","status":"direct model label, unvalidated"}],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2511945,0.001322207,0.004205685,0.04305732,0.002025004,0.007715321,0.003126618,0.002610943,0.004130526],"category_scores_gemma":[0.7149326,0.001589535,0.005120289,0.06376701,0.005396238,0.006235288,0.006377202,0.001885543,0.0006333852],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005037982,"about_ca_system_score_gemma":0.01060438,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002928838,"about_ca_topic_score_gemma":0.00420147,"domain_scores_codex":[0.4200951,0.2347212,0.2304803,0.02590578,0.08648261,0.002315006],"domain_scores_gemma":[0.0945017,0.6658456,0.182054,0.03058521,0.02589663,0.001116875],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001201077,0.0001511034,0.4590096,0.3107244,0.02705315,0.001374988,0.01190522,0.0009583398,0.001720655,0.007678799,0.01170686,0.1665159],"study_design_scores_gemma":[0.0005688589,0.001042092,0.4250259,0.4101196,0.03463173,0.006111494,0.01082034,0.0042583,0.005546734,0.01981913,0.08161414,0.0004417614],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"review","genre_gemma":"empirical","genre_scores_codex":[0.2248424,0.6948489,0.03427274,0.01468204,0.00379995,0.007143433,0.01214088,0.0003111344,0.007958553],"genre_scores_gemma":[0.809733,0.1367808,0.03057097,0.006695218,0.001678638,0.00803548,0.004994187,0.0002303412,0.001281423],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9973891,"threshold_uncertainty_score":0.9234105,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04335486111841079,"score_gpt":0.27953419382027,"score_spread":0.2361793327018592,"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."}}