Cracking the Code: Enhancing Implicit Hate Speech Detection through Coding Classification

2025-06-05

Proceedings of the 5th Workshop on Trustworthy NLP (TrustNLP 2025)

AI Summary

Implicit hate speech (im-HS) poses significant challenges for automated detection due to its semantic opacity and strong contextual dependency. To address this, we propose the first six-dimensional codetype encoding taxonomy, systematically modeling both semantic and rhetorical characteristics of im-HS. Crucially, we integrate this encoding strategy deeply into large language model (LLM) prompt engineering and semantic embedding—marking the first such fusion of codetype guidance with LLM-based inference. We develop a cross-lingual, prompt-driven encoding detection framework grounded in this principle. Evaluated on benchmark Chinese and English im-HS datasets, our method achieves absolute F1-score improvements of 4.2–7.8 percentage points over state-of-the-art baselines. These results robustly demonstrate the effectiveness and cross-lingual generalizability of codetype-guided LLMs for implicit hate speech identification.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search

Abstract

The internet has become a hotspot for hate speech (HS), threatening societal harmony and individual well-being. While automatic detection methods perform well in identifying explicit hate speech (ex-HS), they struggle with more subtle forms, such as implicit hate speech (im-HS). We tackle this problem by introducing a new taxonomy for im-HS detection, defining six encoding strategies named codetypes. We present two methods for integrating codetypes into im-HS detection: 1) prompting large language models (LLMs) directly to classify sentences based on generated responses, and 2) using LLMs as encoders with codetypes embedded during the encoding process. Experiments show that the use of codetypes improves im-HS detection in both Chinese and English datasets, validating the effectiveness of our approach across different languages.

Problem

• Improving detection of subtle implicit hate speech online
• Introducing a new taxonomy with six coding strategies
• Enhancing cross-language performance in hate speech detection

Innovation

• Introducing a new taxonomy for im-HS detection
• Prompting LLMs to classify sentences using codetypes
• Embedding codetypes during LLM encoding process