VLKEB

该数据集是为了评估大型视觉-语言模型（LVLMs）上的知识编辑方法而设立的一个基准，它结合了可靠的数据收集方法和多模态知识图谱。该数据集旨在对LVLMs的编辑可靠性、通用性和局部性进行评估，以推动知识编辑任务的进展。

This dataset is a benchmark established for evaluating knowledge editing methods on Large Vision-Language Models (LVLMs), which combines reliable data collection methods and multimodal knowledge graphs. It aims to assess the editing reliability, generality, and locality of LVLMs, so as to promote the progress of knowledge editing tasks.