JetBrains-Research/synthetic-commit-msg-edits

该数据集是一个基于专家标注的提交信息编辑数据集的合成扩展。数据集包含默认、统计和原始三种配置。默认配置包含由大型语言模型生成的合成信息，统计配置包含指标相关性信息，原始配置用于与可视化应用程序的向后兼容。

This dataset is a synthetic extension based on the expert-labeled commit message edits dataset. The default configuration contains synthetic messages generated by a large language model, with each datapoint including commit hash, repository name, session ID, initial and final commit messages, and data source. The statistics configuration contains information about metrics correlation, while the raw configuration contains all data in a different format for backward compatibility with the visualization app.