HatePrototypes: Interpretable and Transferable Representations for Implicit and Explicit Hate Speech Detection

TL;DR

HatePrototypes introduces class-level vector representations that enable effective, transferable, and parameter-free detection of both explicit and implicit hate speech, reducing the need for repeated fine-tuning.

Contribution

The paper proposes HatePrototypes, a novel approach using class-level vectors derived from language models for efficient, transferable hate speech detection across explicit and implicit categories.

Findings

Prototypes built from 50 examples enable cross-task transfer.

Parameter-free early exiting with prototypes is effective for both hate types.

Code and resources are publicly released for future research.

Abstract

Optimization of offensive content moderation models for different types of hateful messages is typically achieved through continued pre-training or fine-tuning on new hate speech benchmarks. However, existing benchmarks mainly address explicit hate toward protected groups and often overlook implicit or indirect hate, such as demeaning comparisons, calls for exclusion or violence, and subtle discriminatory language that still causes harm. While explicit hate can often be captured through surface features, implicit hate requires deeper, full-model semantic processing. In this work, we question the need for repeated fine-tuning and analyze the role of HatePrototypes, class-level vector representations derived from language models optimized for hate speech detection and safety moderation. We find that these prototypes, built from as few as 50 examples per class, enable cross-task transfer between explicit and implicit hate, with interchangeable prototypes across benchmarks. Moreover, we show that parameter-free early exiting with prototypes is effective for both hate types. We release the code, prototype resources, and evaluation scripts to support future research on efficient and transferable hate speech detection.