StepGame

StepGame数据集是由利兹大学计算机学院创建，用于评估大型语言模型如ChatGPT在空间推理方面的能力。该数据集包含10000个测试样本，每个样本涉及一个故事和相关问题，旨在通过自然语言描述的空间关系来测试模型的推理能力。数据集的创建过程涉及对故事和问题的精确标注，以确保评估的准确性。StepGame数据集主要应用于AI领域，特别是在提高模型对复杂环境理解的能力，以及增强与用户交互的沉浸式体验。

StepGame Dataset was developed by the School of Computer Science, University of Leeds, for evaluating the spatial reasoning capabilities of large language models (LLMs) such as ChatGPT. The dataset contains 10,000 test samples, each consisting of a story and associated questions, aiming to test models' reasoning abilities through spatial relationships described in natural language. The creation process of the StepGame Dataset involves precise annotation of stories and questions to ensure the accuracy of evaluations. The StepGame Dataset is primarily applied in the field of AI, particularly to enhance models' ability to understand complex environments and improve immersive user interaction experiences.