Table 1. The percentage of correct answers for questions in each category.
GPT-3.5 | GPT-4 | p-value | |
Compulsory Questions | 58.0% (29/50) | 90.0% (45/50) | p<0.01 |
General Questions | 64.6% (82/127) | 75.6% (96/127) | p=0.014 |
Scenario-Based Questions | 51.7% (31/60) | 80.0% (48/60) | p<0.01 |
Total Accuracy Rate | 59.9% (142/237) | 79.7% (189/237) | p<0.01 |