GPT-4 can pass the Korean National Licensing Examination for Korean Medicine Doctors
Fig 1
The overall performance of GPT-4 with different prompt designs.
The x-axis and y-axis represent a prompt and the accuracy for the prompt, respectively. The heights of the bars represent the mean of accuracy for multiple trials. A circle mark indicates accuracy in each trial. Dashed lines indicate the chance level of accuracy of 20%, the pass mark of 60%, and the average accuracy of human candidates for the examination of 76.7%.