CodeApex

该数据集是一个双语基准数据集，专注于提升大型语言模型在编程理解和代码生成方面的能力。它包含了多项选择题和算法问题，其中250个多项选择题被分为概念理解、常识推理和多跳推理等类别，同时还包括了476个算法问题，旨在用于代码生成任务。该数据集的任务旨在提高模型在编程理解和代码生成领域的表现。This dataset is a bilingual benchmark dataset focused on advancing the programming comprehension and code generation capabilities of large language models. It encompasses multiple-choice questions and algorithmic problems. Of these, 250 multiple-choice questions are categorized into conceptual understanding, commonsense reasoning, multi-hop reasoning, and other categories. Furthermore, it contains 476 algorithmic problems intended for code generation tasks. The tasks included in this dataset are designed to boost model performance in the domains of programming comprehension and code generation.