JETHICS

该数据集名为JETHICS，是一个用于评估人工智能模型伦理理解能力的日本数据集，包含了78,000个示例，这些示例被分类为规范理论和常识道德两大类。该数据集涵盖了功利主义、义务论、德性伦理学、正义以及常识道德等多个类别，并采用了众包工作者的方式来开发。其规模达到了78,000个示例，旨在对人工智能模型的伦理理解能力进行评估。

This dataset, named JETHICS, is a Japanese benchmark dataset for evaluating the ethical comprehension capabilities of artificial intelligence (AI) models. It contains 78,000 examples categorized into two primary classes: normative theories and commonsense morality. The dataset covers multiple relevant domains including utilitarianism, deontology, virtue ethics, justice, and commonsense morality, and was developed through crowdsourced worker participation. With its 78,000 instances, it aims to assess the ethical understanding abilities of AI models.