OSS-forge/PROBE

PROBE数据集是一个包含编程问题及其解决方案的数据集，以JSONL格式提供。每个条目包含问题ID、自然语言描述、单元测试和参考解决方案。参考解决方案涵盖Python、C++、Java、C和Rust等多种编程语言。数据集基于Project CodeNet，包含来自Aizu和AtCoder平台的问题。数据集主要用于研究和评估大型语言模型在文本到代码生成任务中的表现，支持功能正确性评估、与人类编写解决方案的相似性分析以及代码质量评估。

The PROBE dataset is a collection of programming problems and their solutions, provided in JSONL format. Each entry includes a problem ID, a natural language description, unit tests, and reference solutions. The reference solutions cover multiple programming languages, including Python, C++, Java, C, and Rust. The dataset is based on Project CodeNet and includes problems from the Aizu and AtCoder platforms. It is primarily intended for research and evaluation of Large Language Models in the task of text-to-code generation, supporting functional correctness evaluation, similarity analysis with human-written solutions, and code quality assessment.