ClassEvalQA
收藏arXiv2025-09-30 收录
下载链接：
https://huggingface.co/datasets/datapaf/ClassEvalQABenchmark
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个包含2,050个问题-答案对的基准数据集，这些对子是从类级别的Python编程任务中生成的。问题使用了GPT-3.5进行生成，以确保这些问题在模型预训练期间未曾遇到过。该数据集的任务是对源代码进行问题回答。
This dataset is a benchmark dataset containing 2,050 question-answer pairs generated from class-level Python programming tasks. The questions were created using GPT-3.5 to ensure that they had not been encountered during the model's pre-training stage. The core task of this dataset is question answering over source code.
提供机构：
datapaf
