Curious About Uncertain Scene (CAUS)

CAUS数据集由首尔国立大学和东国大学创建，旨在通过场景描述中的不确定性刺激，促使大型语言模型如GPT-4模仿人类认知过程以解决问题。数据集包含1000个场景描述句子、1000个推理句子和5000个疑问句，全部使用英语。创建过程涉及使用GPT-4生成多样化的场景描述，并通过人工审核确保内容的适宜性和不确定性。CAUS数据集主要应用于人工智能领域，特别是提升AI模型处理不确定性和生成相关问题的能力，从而推动人工智能技术的进步。

The CAUS dataset was developed by Seoul National University and Dongguk University, with the objective of stimulating large language models (LLMs, e.g., GPT-4) to mimic human cognitive processes for problem-solving via uncertainty cues embedded in scene descriptions. It consists of 1000 scene description sentences, 1000 reasoning sentences, and 5000 interrogative sentences, all written in English. The dataset creation process involved using GPT-4 to generate diverse scene descriptions, followed by human review to ensure the content's appropriateness and the presence of uncertainty. The CAUS dataset is primarily applied in the field of artificial intelligence, specifically to enhance AI models' ability to handle uncertainty and generate relevant questions, thereby advancing the development of artificial intelligence technology.