{"id":"W7092191125","doi":"10.1109/ms.2025.3621625","title":"What Inputs Drive Effective Large Language Model-Based Unit Test Generation?","year":2025,"lang":"","type":"article","venue":"IEEE Software","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Unit testing; Correctness; Software quality; Code coverage; Code (set theory); Test (biology); Test Management Approach; Test case","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":["metaepi_narrow","sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.000978462,0.0007448614,0.0006624273,0.0006583928,0.001418025,0.001283608,0.001659018,0.0004996146,0.00002299702],"category_scores_gemma":[0.00309947,0.0007957341,0.0002985653,0.001691894,0.0002171008,0.001314674,0.0006928804,0.0008058812,0.00009494065],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003134489,"about_ca_system_score_gemma":0.001084485,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007520684,"about_ca_topic_score_gemma":0.00006963343,"domain_scores_codex":[0.9957441,0.0003871639,0.0006679462,0.001523722,0.0006465194,0.001030541],"domain_scores_gemma":[0.9938003,0.002715916,0.0003118872,0.002084244,0.0008262089,0.0002613865],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001016179,0.004024118,0.07226044,0.00157985,0.000488294,0.0007804072,0.02977736,0.07953401,0.003838403,0.002553902,0.2003833,0.6046783],"study_design_scores_gemma":[0.001187121,0.0003383288,0.0007485168,0.001877419,0.0001101348,0.000009086434,0.00002429152,0.9567597,0.02845983,0.009024722,0.0005796382,0.0008812257],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02672895,0.002151287,0.9530953,0.0009676892,0.003525529,0.001164737,0.0001091381,0.01216347,0.0000939532],"genre_scores_gemma":[0.8435776,0.00005715512,0.1477549,0.006341422,0.000340568,0.0003138511,0.00005868915,0.00006530317,0.001490591],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8772257,"threshold_uncertainty_score":0.999882,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01701364261331757,"score_gpt":0.2985044136494039,"score_spread":0.2814907710360864,"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."}}