A benchmark dataset for learning to intervene in online hate speech

Abstract

Countering online hate speech is a critical yet challenging task, but one which can be aided by the use of Natural Language Processing (NLP) techniques. Previous research has primarily focused on the development of NLP methods to automatically and effectively detect online hate speech while disregarding further action needed to calm and discourage individuals from using hate speech in the future. In addition, most existing hate speech datasets treat each post as an isolated instance, ignoring the conversational context. In this