Skip to main content
. 2023 Dec 1;2(12):e0000397. doi: 10.1371/journal.pdig.0000397

Fig 3. Accurancy of ChatGPT4 on Chinese National Medical Licensing Examination before encoding.

Fig 3

ChatGPT’s outputs for Units 1, 2, 3, and 4 were evaluated as accurate, inaccurate, or indeterminate using the scoring system outlined in S2 Table after encoding. (A) Assessment of accuracy for open-ended question encodings. (B) Reduced accuracy of encoded questions across Units 1, 2, 3, and 4.