jhu-clsp/astro-llms-benchmark-dataset
收藏Hugging Face2025-07-16 更新2025-08-09 收录
下载链接：
https://hf-mirror.com/datasets/jhu-clsp/astro-llms-benchmark-dataset
下载链接
链接失效反馈官方服务：
资源简介：
AstroLLMs金标基准数据集是一个包含天文学家在一个天文学研究Slack聊天机器人中提出的问题的数据集。每个问题都伴有研究人员团队确定的开放编码标签和专家天文学家的答案。答案要求使用引用，并且不依赖大型语言模型。数据集的结构包括查询、开放编码、答案以及五篇与答案相关的论文。
The AstroLLMs Gold Benchmark Dataset is a collection of queries posed by astronomers to an astronomy research Slack chatbot. Each query is accompanied by open coding labels determined by a team of researchers and answers from expert astronomers. The answers are required to use citations and are not reliant on Large Language Models. The dataset structure includes the query, open coding, answer, and five related papers that pertain to the expert answer.
提供机构：
jhu-clsp
