CONAN

CONAN数据集是由意大利布鲁诺凯斯勒基金会创建的，旨在通过专业文本回应对抗网络仇恨言论。该数据集包含4078对仇恨言论及其对应的反叙事文本，涵盖英语、法语和意大利语。数据收集过程涉及超过100名来自三个不同非政府组织的操作员，他们利用专业训练和经验来完成任务。此外，数据集还提供了关于专家人口统计、仇恨和回应类型的额外注释，以及通过翻译和改写进行的数据增强。CONAN数据集的应用领域主要集中在解决网络仇恨言论问题，通过提供有效的反叙事策略来对抗和缓解仇恨言论的影响。

The CONAN dataset was developed by the Bruno Kessler Foundation in Italy, with the goal of countering online hate speech through professional textual counter-narratives. It consists of 4078 pairs of hate speech and their corresponding counter-narrative texts, covering English, French and Italian languages. The data collection process involved over 100 operators from three distinct non-governmental organizations, who completed the tasks using professional training and practical experience. Additionally, the dataset provides supplementary annotations regarding expert demographics, types of hate speech and their corresponding responses, as well as data augmentation via translation and paraphrasing. The main application fields of the CONAN dataset focus on addressing online hate speech issues, by providing effective counter-narrative strategies to combat and mitigate the impacts of hate speech.