List your company

Hallucination Detection on DiaHalu sampled (test)

245Support (Total Samples)GPT-4
117.08150.29183.5216.71May 12, 2026

Evaluation results

MethodLinks
GPT-4evaluation_category=Overall accuracy2026.05245———54.3
GPT-4evaluation_category=macro avg2026.0524559.554.447.3—
GPT-4evaluation_category=weighted avg2026.0524559.554.347.3—
GPT-4evaluation_category=Class 1 (hallucination)2026.0512366.717.928.2—
GPT-4evaluation_category=Class 0 (factual)2026.0512252.49166.5—