Extended JCommonsenseMorality (eJCM)
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Language-Media-Lab/commonsense-moral-ja
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为eJCM，通过采用一种称为遮蔽标记与标签增强（MTLE）的句子扩展方法，将原始的JCM数据集从13,975个句子扩展至31,184个句子。该数据集特别关注日本文化中的道德推理，并包含了指示句子在道德上是否可接受的标签。规模上，该数据集包含了31,184个句子，其任务是对道德推理和判断进行分类。
This dataset is designated as eJCM. By adopting a sentence expansion approach named Masked Marking and Label Enhancement (MTLE), it expands the original JCM dataset from 13,975 sentences to 31,184 sentences. Focusing specifically on moral reasoning within Japanese culture, this dataset includes labels that indicate whether a sentence is morally acceptable. With a total of 31,184 sentences, it is tailored for the task of classifying moral reasoning and judgments.
提供机构：
Language Media Lab
