SciBench

该数据集包含了一系列大学级别的科学问题，特别侧重于数学相关的问题。在评估中，仅考虑了与数学相关的子集。其难度级别被评定为具有挑战性，任务旨在解决科学问题。This dataset contains a series of university-level scientific problems, with a particular focus on mathematics-related ones. During evaluation, only the mathematics-related subset was considered. Its difficulty level is rated as challenging, and the task aims to solve scientific problems.