{"id":"W4390722673","doi":"10.48550/arxiv.2401.02974","title":"Efficacy of Utilizing Large Language Models to Detect Public Threat Posted Online","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Centennial College","funders":"California State University Long Beach","keywords":"Moderation; Transparency (behavior); Identification (biology); Scale (ratio); Computer security; Internet privacy; Advertising; Computer science; Psychology; Business; Social psychology; Geography","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"],"consensus_categories":[],"category_scores_codex":[0.0003383529,0.0003076546,0.0003916737,0.0006530992,0.0001373966,0.0001081391,0.001758435,0.0002866462,0.00001449506],"category_scores_gemma":[0.00008431587,0.00035311,0.0002598083,0.001496852,0.00003598905,0.0003604912,0.002490701,0.0005074095,0.0001188193],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001583188,"about_ca_system_score_gemma":0.0001755385,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005519406,"about_ca_topic_score_gemma":0.000563793,"domain_scores_codex":[0.9978664,0.0001058812,0.0002615294,0.0010719,0.0001636977,0.0005306132],"domain_scores_gemma":[0.9977736,0.0001191746,0.0002080803,0.001439699,0.0002013324,0.0002580865],"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":[0.0005000035,0.001249047,0.001373159,0.0006967845,0.001056207,0.003149961,0.007870867,0.6734051,0.009907317,0.2543054,0.0008572336,0.04562897],"study_design_scores_gemma":[0.001648051,0.0002604263,0.002483944,0.0003588905,0.0001034906,0.00001219279,0.0007698379,0.9657341,0.005350078,0.02207195,0.0002381259,0.0009689118],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5707253,0.00003925596,0.4274674,0.0001031965,0.0003612742,0.0002986027,0.00005234382,0.0005496842,0.0004028992],"genre_scores_gemma":[0.9954966,0.00006544756,0.003153533,0.00006882281,0.00006090818,0.000001067762,0.00004656928,0.00003312576,0.001073892],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4247713,"threshold_uncertainty_score":0.9998921,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1278831013225399,"score_gpt":0.2267668160144624,"score_spread":0.09888371469192248,"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."}}