{"id":"W4416799470","doi":"10.1109/snpd65828.2025.11252553","title":"Towards Safer Online Platforms: Explainable and Adversarial-Resistant Toxic Comment Detection","year":2025,"lang":"","type":"article","venue":"","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Wilfrid Laurier University","funders":"","keywords":"Adversarial system; SAFER; Filter (signal processing); Vulnerability (computing); Adversarial machine learning; Symbol (formal); Interface (matter)","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":[0.003513755,0.001417831,0.0007194215,0.001371894,0.0007167653,0.001604578,0.00217008,0.002192668,0.002568988],"category_scores_gemma":[0.01802149,0.0003701499,0.0007035094,0.0003890931,0.001178908,0.002594518,0.002527746,0.003018292,0.001942036],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009150957,"about_ca_system_score_gemma":0.001153063,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002275032,"about_ca_topic_score_gemma":0.003273724,"domain_scores_codex":[0.9971559,0.001215667,0.0001087024,0.0005527561,0.0007709998,0.0001959787],"domain_scores_gemma":[0.9898331,0.005517545,0.001270644,0.00182273,0.001222287,0.000333729],"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":[0.001065626,0.0007760595,0.03943186,0.0006507476,0.0002052814,0.001450737,0.002173101,0.2131887,0.03172461,0.02153176,0.04524621,0.6425553],"study_design_scores_gemma":[0.00001743924,0.00006338728,0.001431149,0.00003715346,0.00001481483,0.0001483816,0.0001305553,0.9771548,0.007766151,0.01003055,0.003181291,0.00002433959],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1379668,0.0007969297,0.825296,0.003366888,0.000286759,0.0004862745,0.001363173,0.02431468,0.006122384],"genre_scores_gemma":[0.8075786,0.0002096598,0.1838188,0.0006947392,0.0001768286,0.0001706191,0.001643235,0.0004481366,0.005259301],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003513755,"threshold_uncertainty_score":0.0185827,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01211647022087386,"score_gpt":0.2427711181796792,"score_spread":0.2306546479588054,"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."}}