osunlp/AutoSDT-5K

AutoSDT-5K是一个包含5,404个编码任务的自动构建数据集，服务于数据驱动的科学发现，涵盖四个科学领域，使用了756个独特的Python包。该数据集是迄今为止最大的开放数据集，专门用于数据驱动的科学发现。AutoSDT-5K is an automatically constructed dataset of 5,404 coding tasks for data-driven discovery, covering four scientific disciplines and utilizing 756 unique Python packages. It is the largest open dataset to date, specifically for data-driven scientific discovery.