Detecting Hate Speech with GPT-3
Ke-Li Chiu University of Toronto Annie Collins University of Toronto Rohan Alexander University of Toronto and Schwartz Reisman Institute
Abstract
Sophisticated language models such as OpenAI’s GPT-3 can generate hateful text that targets marginalized groups. Given this capacity, we are interested in whether large language models can be used to identify hate speech and classify text as sexist or racist. We use GPT-3 to identify sexist and racist text passages with zero-, one-, and few-shot learning. We find that with zero- and one-shot learning, GPT-3 can identify sexist or racist text with an average accuracy between 55 per cent and 67 per cent, depending on the category of text and type of learning. With few-shot learning, the model’s accuracy can be as high as 85 per cent. Large language models have a role to play in hate speech detection, and with further development they could eventually be used to counter hate speech.
中文速览
大型语言模型（如GPT-3）既能生成仇恨言论，研究者就想知道它能否反过来用于识别这类有害内容。作者从ETHOS数据集中抽取120至243条被标注为种族歧视、性别歧视或中性的真实评论，分别用零样本、单样本和少样本（few-shot learning）三种方式提示GPT-3进行分类。结果显示，零样本和单样本学习下平均准确率在55%至67%之间，而混合类别的少样本学习可将准确率提升至约85%，F1分数达到77%至79%。这项研究说明，大型语言模型凭借对仇恨话语模式的"内隐知识"，有望在仇恨言论自动检测乃至自我审查方面发挥实际作用，为互联网内容治理提供了一条低标注成本的新路径。
原文 arXiv:2103.12407;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2103.12407v4