ChatGPT Role-play Dataset (CRD)

ChatGPT角色扮演数据集（CRD）是由波特兰州立大学计算机科学系的研究团队创建的，包含85个独特的对话，总计1742条语句。该数据集通过手动标注用户动机和模型自然度，旨在探索人类与ChatGPT在正常和角色扮演对话场景中的互动。CRD数据集分为三个子集：普通对话、老板角色扮演和同学角色扮演，每个子集都旨在研究不同的社交和语言行为。该数据集的应用领域包括改进人机交互的自然性和效率，以及探索AI在模拟人类对话中的潜力。

ChatGPT Role-Playing Dataset (CRD) was developed by a research team from the Department of Computer Science, Portland State University. It contains 85 unique conversations, totaling 1,742 utterances. With manual annotations of user motivations and model naturalness, this dataset is designed to explore human-ChatGPT interactions in both normal and role-playing dialogue scenarios. The CRD dataset is split into three subsets: general conversation, boss role-play, and classmate role-play, each targeting the investigation of distinct social and linguistic behaviors. Potential applications of this dataset include improving the naturalness and efficiency of human-computer interaction, as well as exploring the potential of AI in simulating human conversations.