Alongside the dataset, we further conducted thorough experiments with representative LLMs and QA algorithms on CMExam. Excluding non-textual questions may introduce biases. BLEU and ROUGE metrics are ...