BIRD-CRITIC-Multi
收藏arXiv2025-09-30 收录
下载链接：
https://bird-critic.github.io/
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了570个多方言任务的基准测试集，这些任务源自真实的用户问题，旨在评估在不同SQL方言中处理SQL问题调试的能力。此外，该基准测试集中的任务能够评估大型语言模型在不同SQL方言问题上的性能表现。任务类型为SQL问题调试。
This dataset is a benchmark test set comprising 570 multi-dialect tasks. Derived from real user queries, these tasks are designed to evaluate the capability of debugging SQL problems across different SQL dialects. Furthermore, the tasks in this benchmark can assess the performance of large language models (LLMs) on issues involving various SQL dialects. All tasks fall under the category of SQL problem debugging.
