Blanca/MAFALDA-fallacies

MAFALDA是一个统一的谬误分类数据集，它合并和统一了之前的谬误数据集，并带有一种新的分类法，以对现有的谬误分类进行对齐、优化和统一。数据集的一部分还提供了人工注释和每个注释的人工解释。此外，该数据集提出了一个新的针对主观NLP任务的注释方案，以及一种新的评估方法，旨在处理主观性。

MAFALDA is a unified fallacy classification dataset that merges and unifies previous fallacy datasets, coming with a new taxonomy that aligns, refines, and unifies existing classifications of fallacies. A part of the dataset is provided with manual annotations and explanations for each annotation. Moreover, the dataset introduces a new annotation scheme tailored for subjective NLP tasks, as well as a new evaluation method designed to handle subjectivity.