{"id":"W7118191471","doi":"10.1109/aiccsa66935.2025.11315450","title":"Ensemble Hallucination Detection in Arabic Large Language Models: A Multi-Modal Linguistic and Self-Consistency Approach","year":2025,"lang":"","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université du Québec en Abitibi-Témiscamingue","funders":"","keywords":"Readability; Arabic; Heuristics; Feature (linguistics); Software deployment","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":[0.002591529,0.001237339,0.0008027198,0.00105391,0.0005395933,0.001199441,0.001172141,0.0006953792,0.001012673],"category_scores_gemma":[0.008522578,0.0002909609,0.0008128192,0.0004930136,0.0003861628,0.00182166,0.002059784,0.001474681,0.0009486633],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004184786,"about_ca_system_score_gemma":0.0007244584,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00380766,"about_ca_topic_score_gemma":0.005459826,"domain_scores_codex":[0.9990151,0.0003752626,0.00007075797,0.0002577185,0.0001818699,0.00009926595],"domain_scores_gemma":[0.9965364,0.001643722,0.0002459876,0.0006001783,0.0007675432,0.000206186],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001028613,0.0004317319,0.02431922,0.0002189355,0.0004554504,0.0005084479,0.001076957,0.1246256,0.04425908,0.002387173,0.009094027,0.7915947],"study_design_scores_gemma":[0.00001489977,0.00009150172,0.001721604,0.00001248368,0.00005113155,0.0001208244,0.0002112002,0.9844726,0.009942225,0.002364109,0.0009757665,0.00002169213],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.310012,0.000938935,0.6738809,0.0006409757,0.0001299217,0.0001729906,0.0006798144,0.01155111,0.001993414],"genre_scores_gemma":[0.8744391,0.0001683727,0.1216664,0.000185319,0.00006817519,0.00007734343,0.001662651,0.0003578941,0.001374712],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00380766,"threshold_uncertainty_score":0.01370549,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02053379864611973,"score_gpt":0.2602422438399967,"score_spread":0.2397084451938769,"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."}}