VisEval
收藏资源简介:
VisEval是一个用于自然语言到可视化(NL2VIS)任务的高质量大型数据集,由微软研究院主导创建。该数据集包含2,524个查询,涉及146个数据库,旨在评估大型语言模型在可视化生成方面的能力。数据集通过结合先进的LLMs智能和可视化专家经验进行筛选和标注,确保了查询的高质量和准确性。VisEval数据集主要应用于评估和提升LLMs在数据可视化领域的性能,特别是在自然语言处理和可视化设计方面。
VisEval is a high-quality large-scale dataset for the natural language to visualization (NL2VIS) task, developed and led by Microsoft Research. It contains 2,524 queries covering 146 databases, aiming to evaluate the capabilities of large language models (LLMs) in visualization generation. The dataset is curated and annotated by integrating the advanced capabilities of LLMs and the professional expertise of visualization specialists, ensuring the high quality and accuracy of the queries. The VisEval dataset is primarily utilized to evaluate and enhance the performance of LLMs in the domain of data visualization, particularly in natural language processing and visualization design.




