SYSUSELab/MultiCodeIF
收藏Hugging Face2025-07-02 更新2025-10-18 收录
下载链接：
https://hf-mirror.com/datasets/SYSUSELab/MultiCodeIF
下载链接
链接失效反馈官方服务：
资源简介：
MultiCodeIF是一个自动构建的基准数据集，用于评估代码生成中的指令遵循能力，涵盖14种编程语言的2021个任务，包含功能性和非功能性的多层次约束，支持多种类型、多层次、多轮次的评估。
MultiCodeIF is an automatically constructed benchmark for evaluating instruction-following ability in code generation, covering 2021 tasks across 14 programming languages, with multi-level constraints for both functional and non-functional aspects, supporting multi-type, multi-level, and multi-turn evaluation settings.
提供机构：
SYSUSELab
