抗议事件数据集
收藏资源简介:
抗议事件数据集由莱顿大学和迭戈波塔莱斯大学创建,包含超过350万条关于阿根廷和智利抗议事件的推文。数据集涵盖了2020年阿根廷的冠状病毒和司法改革抗议以及2019年智利的社会爆发抗议。数据集通过Twitter API获取,并使用Perspective算法和人工标注创建了黄金标准,用于评估大型语言模型在文本标注任务中的表现。该数据集主要用于研究社交媒体中的毒性和不文明行为,旨在解决政治内容自动标注的问题。
The Protest Event Dataset, created by Leiden University and Universidad Diego Portales, contains over 3.5 million Tweets related to protest events in Argentina and Chile. It covers two major protest campaigns: the 2020 protests against COVID-19 and judicial reforms in Argentina, and the 2019 social uprising protests in Chile. The dataset was collected via the Twitter API, and a gold standard for evaluating the performance of Large Language Models (LLMs) on text annotation tasks was developed using the Perspective algorithm and manual annotations. This dataset is primarily used for research on toxicity and incivility in social media, aiming to address the challenges of automatic annotation for political content.

- 1Benchmarking LLMs in Political Content Text-Annotation: Proof-of-Concept with Toxicity and Incivility Data莱顿大学, 荷兰 迭戈波塔莱斯大学, 智利 · 2024年



