SITT Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/social-influence/sitt-dataset/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含746个对话，由11位专家进行了标注，旨在识别文本内容中微妙的社交影响形式。这些专家共完成了2177项标注任务，涵盖了58种技巧，这些技巧被组织成了九个类别。数据集的规模为746个对话，平均每个对话包含6.46个回合。这项任务的目标是对对话中的社交影响技巧进行检测和分类。
This dataset includes 746 dialogues, which were annotated by 11 experts for the purpose of identifying subtle forms of social influence in textual content. A total of 2177 annotation tasks were completed by these experts, covering 58 social influence techniques that are grouped into nine categories. The dataset consists of 746 dialogues, with an average of 6.46 conversational turns per dialogue. The core objective of this task is to detect and classify the social influence techniques within dialogues.
提供机构：
Social Influence Project
