bench-temp/mmlu-pt-high-school-ablation-v2-notink-natural
收藏数据链接：
官方服务：
资源简介：
该数据集包含多个学科领域的任务难度评估数据，特征包括任务原始文本、问题ID、正确数难度和分数难度，按健康与生物科学、人文学科、语言、STEM四个领域分割。
This dataset contains task difficulty evaluation data across multiple academic domains, with features including task raw text, question ID, difficulty based on number of correct answers, and difficulty score. It is split into four domains: health and biological sciences, humanities, languages, and STEM.
提供机构：
bench-temp
