smy-hit/TopoTalk-Bench

该数据集旨在测试大型语言模型处理网络拓扑的能力，包括两个主要任务：从零开始构建和修改现有拓扑。数据集包含原始网络拓扑（空和原始）、自然语言描述（中文和英文）以及地面真实NetJSON文件。总共有1278个文件，分为五种类型（A-E），分别对应不同的操作（如构建、添加/删除节点/链接）。每种类型和语言都有详细的统计信息。数据集还提供了文件命名约定、节点类型参考以及生成信息。

This dataset is designed to test large language models ability to process network topologies, including two main tasks: building from scratch and modifying existing topologies. It contains original network topologies (empty and original), natural language descriptions (Chinese and English), and ground truth NetJSON files. There are a total of 1278 files, categorized into five types (A-E) corresponding to different operations (e.g., building, adding/deleting nodes/links). Detailed statistics are provided for each type and language. The dataset also includes file naming conventions, node type references, and generation information.