{"id":"W4416799210","doi":"10.1109/snpd65828.2025.11252591","title":"Making the Case for LLM-Generated Automated Program Repair Benchmarks","year":2025,"lang":"","type":"article","venue":"","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Algoma University","funders":"","keywords":"Overfitting; Benchmark (surveying); Benchmarking; Focus (optics); Quality (philosophy); Quality assurance","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":[0.026681,0.001153113,0.000697921,0.002368164,0.0007739784,0.003282017,0.003668345,0.001832379,0.001791898],"category_scores_gemma":[0.189551,0.0005765716,0.0006616949,0.001956621,0.001945788,0.004565858,0.003399949,0.003588692,0.00106517],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001798524,"about_ca_system_score_gemma":0.003232607,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00300817,"about_ca_topic_score_gemma":0.006483141,"domain_scores_codex":[0.9684805,0.02036177,0.001785262,0.00234527,0.006338808,0.0006885293],"domain_scores_gemma":[0.8848275,0.064278,0.005410044,0.027535,0.01634674,0.00160276],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001575358,0.001530102,0.04666546,0.002297729,0.0003918201,0.0005645339,0.001643969,0.2103869,0.02232483,0.08633135,0.1282212,0.4980668],"study_design_scores_gemma":[0.0005957386,0.001258313,0.01082405,0.001194146,0.0001359405,0.0003571408,0.0008574432,0.7490356,0.03401427,0.1178332,0.08374391,0.0001502387],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.2431432,0.004848833,0.6525882,0.02113072,0.002088932,0.001325912,0.007053354,0.04154371,0.02627722],"genre_scores_gemma":[0.5088752,0.0006173532,0.4687972,0.004097927,0.0001921051,0.001423784,0.01052845,0.0034221,0.002045824],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.026681,"threshold_uncertainty_score":0.1411043,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04917975613988371,"score_gpt":0.3669347614918263,"score_spread":0.3177550053519426,"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."}}