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

arXiv: https://arxiv.org/abs/2511.06391 · HTML: https://mlsanity.com/paper/2511.06391/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 51 references that are in the map.
- First posted on arXiv 9 Nov 2025.

## Builds on (most cited first, 27 of 27)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Prototypical Networks for Few-shot Learning](https://mlsanity.com/paper/1703.05175/) (2017)
- [Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations](https://mlsanity.com/paper/2312.06674/) (2023)
- [StereoSet: Measuring stereotypical bias in pretrained language models](https://mlsanity.com/paper/2004.09456/) (2020)
- [CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models](https://mlsanity.com/paper/2010.00133/) (2020)
- [ToxiGen: A Large-Scale Machine-Generated Dataset for Adversarial and Implicit Hate Speech Detection](https://mlsanity.com/paper/2203.09509/) (2022)
- [Social Bias Frames: Reasoning about Social and Power Implications of Language](https://mlsanity.com/paper/1911.03891/) (2019)
- [DeeBERT: Dynamic Early Exiting for Accelerating BERT Inference](https://mlsanity.com/paper/2004.12993/) (2020)
- [LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding](https://mlsanity.com/paper/2404.16710/) (2024)
- [BERT Loses Patience: Fast and Robust Inference with Early Exit](https://mlsanity.com/paper/2006.04152/) (2020)
- [Predicting the Type and Target of Offensive Posts in Social Media](https://mlsanity.com/paper/1902.09666/) (2019)
- [FastBERT: a Self-distilling BERT with Adaptive Inference Time](https://mlsanity.com/paper/2004.02178/) (2020)
- [Latent Hatred: A Benchmark for Understanding Implicit Hate Speech](https://mlsanity.com/paper/2109.05322/) (2021)
- [HateCheck: Functional Tests for Hate Speech Detection Models](https://mlsanity.com/paper/2012.15606/) (2020)
- [A New Generation of Perspective API: Efficient Multilingual Character-level Transformers](https://mlsanity.com/paper/2202.11176/) (2022)
- [Directions in Abusive Language Training Data: Garbage In, Garbage Out](https://mlsanity.com/paper/2004.01670/) (2020)
- [Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding](https://mlsanity.com/paper/2310.05424/) (2023)
- [Efficient Methods for Natural Language Processing: A Survey](https://mlsanity.com/paper/2209.00099/) (2022)
- [Text Detoxification using Large Pre-trained Neural Models](https://mlsanity.com/paper/2109.08914/) (2021)
- [Explicit vs. Implicit: Investigating Social Bias in Large Language Models through Self-Reflection](https://mlsanity.com/paper/2501.02295/) (2025)
- [Hate Speech Criteria: A Modular Approach to Task-Specific Hate Speech Definitions](https://mlsanity.com/paper/2206.15455/) (2022)
- [HateDay: Insights from a Global Hate Speech Dataset Representative of a Day on Twitter](https://mlsanity.com/paper/2411.15462/) (2024)
- [DEED: Dynamic Early Exit on Decoder for Accelerating Encoder-Decoder Transformer Models](https://mlsanity.com/paper/2311.08623/) (2023)
- [Towards Detecting Contextual Real-Time Toxicity for In-Game Chat](https://mlsanity.com/paper/2310.18330/) (2023)
- [DE3-BERT: Distance-Enhanced Early Exiting for BERT based on Prototypical Networks](https://mlsanity.com/paper/2402.05948/) (2024)
- [Noisy Self-Training with Data Augmentations for Offensive and Hate Speech Detection Tasks](https://mlsanity.com/paper/2307.16609/) (2023)

## Abstract (arXiv)

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.