OpenTWBench/

tw-chinese-language-bench

A multiple-choice benchmark for Chinese Language and Literature — the subject 「國文」 as examined in Taiwan's national examinations, in Traditional Chinese. Built from the papers published as open data by the Ministry of Examination (考選部). One dataset per academic subject: papers are grouped by the subject they examine, not by the professional category sitting the exam. Part of OpenTWBench — an open evaluation suite for Taiwan-domain knowledge.… See the full description on the dataset page: https://huggingface.co/datasets/OpenTWBench/tw-chinese-language-bench.