ResearchBenchmarksConfidence Estimation on MedQA (test)Follow0.69AUROCMedConf0.640080.653040.6660.67896Jan 22, 2026Evaluation ResultsMethodMethodLinksAUROCPearson CorrelationMedConfType=Self-Verbalized,...Type=Self-Verbalized, Model=GPT-4.12026.010.690.989EccType=Consistency-Level...Type=Consistency-Level, Model=GPT-4.12026.010.6670.906CoT CEType=Self-Verbalized,...Type=Self-Verbalized, Model=GPT-4.12026.010.6420.987