{"id":"W4404782839","doi":"10.18653/v1/2024.emnlp-main.243","title":"STOP! Benchmarking Large Language Models with Sensitivity Testing on Offensive Progressions","year":2024,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Brock University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Offensive; Benchmarking; Computer science; Sensitivity (control systems); Reliability engineering; Engineering; Operations research; Electronic engineering; Business","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01251768,0.003480415,0.001334563,0.00241088,0.001215424,0.003105555,0.003099242,0.002874034,0.006235002],"category_scores_gemma":[0.06058193,0.0008183814,0.002189418,0.001780768,0.001390425,0.005648965,0.003454815,0.004873519,0.004736573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001518817,"about_ca_system_score_gemma":0.002254416,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01583823,"about_ca_topic_score_gemma":0.02221315,"domain_scores_codex":[0.9886881,0.006509674,0.0007969163,0.002131546,0.001354995,0.0005186884],"domain_scores_gemma":[0.9617374,0.02874685,0.0007600089,0.005102768,0.002866745,0.0007862863],"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":[0.003055806,0.001852238,0.04948543,0.00237328,0.002115905,0.001091225,0.001894824,0.437045,0.01163797,0.009275074,0.1744932,0.3056801],"study_design_scores_gemma":[0.0003499083,0.0004402772,0.004217658,0.0001734607,0.0001638072,0.0002842047,0.0005142394,0.9582462,0.007419497,0.01379611,0.01426722,0.0001274185],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5994344,0.007263386,0.2309149,0.004491362,0.002370965,0.001256949,0.04180084,0.09243698,0.02003019],"genre_scores_gemma":[0.7989014,0.0007313564,0.1074027,0.001801203,0.0003731785,0.0007562687,0.08236494,0.004102107,0.003567004],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01583823,"threshold_uncertainty_score":0.06620061,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03257478070019768,"score_gpt":0.2729869520796337,"score_spread":0.2404121713794361,"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."}}