{"id":"W4416805858","doi":"10.2139/ssrn.5824882","title":"Scientific Workflow Development Using Large Language Models","year":2025,"lang":"","type":"preprint","venue":"SSRN Electronic Journal","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Workflow; Workflow technology; Workflow engine; Subject-matter expert; Domain (mathematical analysis); Pipeline (software); Correctness","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004517881,0.0007619882,0.000727722,0.001131201,0.001148239,0.004206111,0.001881993,0.001075404,0.005033825],"category_scores_gemma":[0.01961775,0.00117152,0.003217428,0.001078613,0.0007655975,0.004885807,0.002408614,0.002812075,0.001958947],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001568216,"about_ca_system_score_gemma":0.00457886,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008027468,"about_ca_topic_score_gemma":0.01556373,"domain_scores_codex":[0.9967638,0.001536148,0.000268598,0.000470711,0.0007883857,0.0001723632],"domain_scores_gemma":[0.981682,0.01357532,0.0006518807,0.002360601,0.001363727,0.0003665806],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0004655922,0.0004840601,0.00490694,0.0004571924,0.0003601585,0.0007233825,0.001061724,0.629773,0.00727999,0.2103281,0.01045833,0.1337016],"study_design_scores_gemma":[0.00002089663,0.00002296767,0.00009124801,0.00001876312,0.00003439681,0.00003258728,0.00004179161,0.9442402,0.00196634,0.04977167,0.003744924,0.00001426413],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00952664,0.00005314741,0.9833466,0.0003487953,0.00003279036,0.00009076644,0.0004127572,0.004256341,0.00193221],"genre_scores_gemma":[0.2712625,0.0002255075,0.7199822,0.0002121694,0.00004235884,0.0003945494,0.001974678,0.001589052,0.004316968],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008027468,"threshold_uncertainty_score":0.02389312,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07686589203142266,"score_gpt":0.3636975414047908,"score_spread":0.2868316493733681,"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."}}