ResearchBenchmarksBinary classification (Human vs Machine speech) on MultiDialog (Human-Human) OOD (test)Binary classification (Human vs Machine speech) on MultiDialog (Human-Human) OOD (test)Follow95.31Accuracyinterpretable AI judge90.544592.9272595.3197.69275Feb 27, 2026Evaluation resultsMethodMethodLinksAccuracyinterpretable AI judge2026.0295.31