CausalTalk
收藏资源简介:
CausalTalk是一个多层次的Reddit帖子数据集,包含2020年至2024年间关于COVID-19大流行相关公共卫生的讨论。数据集涵盖了四个任务:二元因果分类、显式与隐式因果性检测、原因-结果跨度提取和因果概要生成。数据集包括专家手动标注的金标准标签和通过GPT-4o生成的银标准标签。CausalTalk旨在支持细粒度因果检测和基于概要的推理,为研究社交媒体中的因果推理提供了丰富的资源。
CausalTalk is a multi-level Reddit post dataset containing discussions related to public health concerning the COVID-19 pandemic spanning from 2020 to 2024. The dataset covers four tasks: binary causal classification, explicit and implicit causality detection, cause-effect span extraction, and causal summary generation. It includes gold-standard labels manually annotated by experts and silver-standard labels generated by GPT-4o. CausalTalk aims to support fine-grained causal detection and summary-based reasoning, providing a rich resource for research on causal reasoning in social media.
CausalTalk数据集概述
数据集名称
CausalTalk
数据集来源
GitHub仓库:https://github.com/xding2/CausalTalk




