LingoIITGN/PythonSaga
收藏Hugging Face2025-03-17 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/LingoIITGN/PythonSaga
下载链接
链接失效反馈官方服务：
资源简介：
PythonSaga是一个针对Python代码生成的大型语言模型（LLM）评估数据集，包含了185个手工制作的提示，这些提示平衡地代表了38个编程概念，并涵盖了不同的难度级别。该数据集旨在改进代码生成语言模型评估基准，通过展示现有模型在该数据集上的性能，揭示了PythonSaga在评估代码生成模型方面的有效性和挑战性。
PythonSaga is an evaluation dataset for large language models (LLMs) focused on Python code generation, consisting of 185 hand-crafted prompts that represent 38 programming concepts across various difficulty levels in a balanced manner. The dataset aims to improve the benchmarks for evaluating code generation language models by demonstrating the performance of existing models on it, revealing the effectiveness and challenges of PythonSaga in assessing code generation models.
提供机构：
LingoIITGN
