{"id":"W4399280155","doi":"10.1101/2024.06.01.24308323","title":"Prompting is all you need: LLMs for systematic review screening","year":2024,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":16,"is_retracted":false,"has_abstract":true,"ca_institutions":"Health Sciences Centre; University of Calgary; Sunnybrook Health Science Centre; Vector Institute; South Health Campus; University of Toronto","funders":"","keywords":"Medicine","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":["metaresearch","metaepi_narrow","scholarly_communication","insufficient_payload"],"consensus_categories":["metaresearch","insufficient_payload"],"category_scores_codex":[0.2537764,0.0009523713,0.01399967,0.0006717514,0.0001545428,0.002697273,0.005197378,0.0003220794,0.003801002],"category_scores_gemma":[0.09365731,0.0004390554,0.008231882,0.001297126,0.00004626417,0.0001081685,0.001990461,0.0008117582,0.007646489],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000602947,"about_ca_system_score_gemma":0.0001405892,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001316257,"about_ca_topic_score_gemma":0.000005846905,"domain_scores_codex":[0.9568931,0.008903673,0.02273807,0.00295774,0.007905583,0.0006018125],"domain_scores_gemma":[0.9695544,0.005474637,0.01300844,0.009731473,0.001929721,0.000301386],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.000001535278,0.00001985402,0.00046804,0.8302141,0.001574179,0.00001217191,0.0006618638,0.00004041615,0.00001155663,0.0004142225,0.1651477,0.001434316],"study_design_scores_gemma":[0.0001810451,0.00005341985,0.00009534742,0.7879292,0.01830055,0.00006337086,0.000862065,0.06899628,0.00003357308,0.02319849,0.09887026,0.001416446],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"other","genre_scores_codex":[0.01505473,0.7496055,0.1401827,0.02363525,0.004584233,0.0534999,0.0005091965,0.0001925141,0.01273601],"genre_scores_gemma":[0.1991618,0.01543461,0.2979062,0.03532489,0.004589105,0.03925075,0.0004918779,0.001168539,0.4066722],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.7341709,"threshold_uncertainty_score":0.9998061,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.741785538676171,"score_gpt":0.5428035855572323,"score_spread":0.1989819531189387,"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."}}