Can Prompting LLMs Unlock Hate Speech Detection across Languages? A Zero-shot and Few-shot Study
This addresses the challenge of linguistic diversity in online content for researchers and practitioners in NLP, but it is incremental as it builds on existing prompting techniques.
The study tackled the problem of hate speech detection across multiple languages by evaluating zero-shot and few-shot prompting of LLMs, finding that while these methods lag behind fine-tuned models on real-world datasets, they achieve better generalization on functional tests.
Despite growing interest in automated hate speech detection, most existing approaches overlook the linguistic diversity of online content. Multilingual instruction-tuned large language models such as LLaMA, Aya, Qwen, and BloomZ offer promising capabilities across languages, but their effectiveness in identifying hate speech through zero-shot and few-shot prompting remains underexplored. This work evaluates LLM prompting-based detection across eight non-English languages, utilizing several prompting techniques and comparing them to fine-tuned encoder models. We show that while zero-shot and few-shot prompting lag behind fine-tuned encoder models on most of the real-world evaluation sets, they achieve better generalization on functional tests for hate speech detection. Our study also reveals that prompt design plays a critical role, with each language often requiring customized prompting techniques to maximize performance.