grow-ai-like-a-child/intentionality

这是一个包含图像-问题对的视角采纳任务数据集，共有100个样本，分为训练集和测试集。每个样本包括一个或多个图像文件、一个关于图像的问题以及一个答案（如果有）。数据集按照概念类别组织，每个样本都有一个问题标识符、概念类别、问题文本、答案文本（如有）、主图像文件名、附加图像（如有）、附加文本文件内容和训练或测试集标识。This dataset contains image-question pairs for perspective-taking tasks, with a total of 100 samples divided into training and testing sets. Each sample includes one or more image files, a question about the image, and an answer (if available). The dataset is organized by concept categories, with each sample having a question identifier, concept category, question text, answer text (if available), primary image filename, additional images (if any), additional text file content, and training or test set label.