{"id":"W4407986251","doi":"10.1007/s10664-025-10619-z","title":"An empirical study on LLM-based classification of requirements-related provisions in food-safety regulations","year":2025,"lang":"en","type":"article","venue":"Empirical Software Engineering","topic":"Technology and Data Analysis","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Ottawa","funders":"Natural Sciences and Engineering Research Council of Canada; Ontario Centre of Innovation","keywords":"Food safety; Empirical research; Business; Computer science; Food science; Chemistry; Statistics; Mathematics","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.0003761635,0.0001521755,0.0002518922,0.0006684231,0.00009189205,0.00002806277,0.0006792341,0.0001672407,0.000007152832],"category_scores_gemma":[0.0006444785,0.0001457182,0.00006961461,0.002392478,0.00003318385,0.000262034,0.0001179595,0.0002950316,0.000005775739],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001379295,"about_ca_system_score_gemma":0.00009503092,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000008518087,"about_ca_topic_score_gemma":0.00001870708,"domain_scores_codex":[0.9984426,0.00009217099,0.0005227199,0.0004785557,0.000245161,0.0002187772],"domain_scores_gemma":[0.9984913,0.0003442553,0.00007718233,0.0009614989,0.00006420905,0.00006158406],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00002261174,0.002286624,0.9073421,0.00002270035,0.00009219996,0.000007470446,0.0004429429,0.07648394,0.0004444764,0.005535738,0.0003006097,0.007018585],"study_design_scores_gemma":[0.0004624904,0.0003516178,0.7975535,0.00004303358,0.00001541928,2.331796e-7,0.0000316771,0.2003697,0.0002316581,0.000401501,0.0004169032,0.0001221821],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3844881,0.00001516813,0.6138507,0.0009525428,0.00007129531,0.0002104131,0.000006234636,0.0003817513,0.00002389091],"genre_scores_gemma":[0.984479,4.908318e-7,0.01533288,0.00008545232,0.000005564064,0.00004080568,0.00002974691,0.000007848167,0.00001819863],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.599991,"threshold_uncertainty_score":0.5942218,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03242033749613746,"score_gpt":0.3338891427159063,"score_spread":0.3014688052197688,"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."}}