{"id":"W4411449748","doi":"10.1145/3715735","title":"Hallucination Detection in Large Language Models with Metamorphic Relations","year":2025,"lang":"en","type":"article","venue":"Proceedings of the ACM on software engineering.","topic":"Topic Modeling","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University; University of Calgary","funders":"","keywords":"Recall; Computer science; Margin (machine learning); Cognitive psychology; Psychology; Artificial intelligence; Machine learning","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":[],"consensus_categories":[],"category_scores_codex":[0.0002345763,0.00009569502,0.0001103577,0.0002396517,0.00004339164,0.00003424868,0.0009122883,0.00005178755,8.93768e-7],"category_scores_gemma":[0.0007983778,0.00007314536,0.00003282861,0.0006994802,0.000006595974,0.0003993611,0.0003016699,0.0001885808,0.000001154058],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007704931,"about_ca_system_score_gemma":0.00002016248,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001483219,"about_ca_topic_score_gemma":0.000009156996,"domain_scores_codex":[0.9992983,0.000002920567,0.0001578305,0.0002095511,0.0001873912,0.0001440128],"domain_scores_gemma":[0.9994018,0.00005874653,0.00006502929,0.000365574,0.00009236373,0.00001647049],"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":[0.00003888557,0.0002125209,0.01937726,0.0004122273,0.00009680699,0.000002345696,0.004469331,0.599224,0.008829392,0.3446894,0.000165616,0.02248222],"study_design_scores_gemma":[0.0004130699,0.00003208543,0.01438171,0.0002458688,0.00001268524,0.000003027385,0.00004000212,0.9647879,0.01565256,0.004211365,0.00008794199,0.0001317691],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2859885,0.00006599667,0.7129685,0.0003804779,0.00009912989,0.0001488211,7.859352e-7,0.0001723813,0.0001754253],"genre_scores_gemma":[0.8660687,0.000001405243,0.1336897,0.00004641654,0.0000100873,0.00002610524,2.067285e-7,0.00000704162,0.0001503352],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5800802,"threshold_uncertainty_score":0.2982781,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00800568433846788,"score_gpt":0.2089088547174961,"score_spread":0.2009031703790282,"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."}}