Towards Efficient and Explainable Hate Speech Detection via Model Distillation
This work addresses the need for efficient and explainable hate speech detection for online platforms, though it is incremental as it builds on existing distillation and explanation methods.
The paper tackled the problem of computationally expensive and non-interpretable hate speech detection by distilling large language models using Chain-of-Thought to extract explanations, resulting in distilled models that match explanation quality and surpass classification performance of larger models.
Automatic detection of hate and abusive language is essential to combat its online spread. Moreover, recognising and explaining hate speech serves to educate people about its negative effects. However, most current detection models operate as black boxes, lacking interpretability and explainability. In this context, Large Language Models (LLMs) have proven effective for hate speech detection and to promote interpretability. Nevertheless, they are computationally costly to run. In this work, we propose distilling big language models by using Chain-of-Thought to extract explanations that support the hate speech classification task. Having small language models for these tasks will contribute to their use in operational settings. In this paper, we demonstrate that distilled models deliver explanations of the same quality as larger models while surpassing them in classification performance. This dual capability, classifying and explaining, advances hate speech detection making it more affordable, understandable and actionable.