# Analyzing Bias in False Refusal Behavior of Large Language Models for Hate Speech Detoxification

arXiv: https://arxiv.org/abs/2601.08668 · HTML: https://mlsanity.com/paper/2601.08668/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 45 references that are in the map.
- First posted on arXiv 13 Jan 2026.

## Builds on (most cited first, 25 of 25)

- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Jailbroken: How Does LLM Safety Training Fail?](https://mlsanity.com/paper/2307.02483/) (2023)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judge](https://mlsanity.com/paper/2411.16594/) (2024)
- [HateXplain: A Benchmark Dataset for Explainable Hate Speech Detection](https://mlsanity.com/paper/2012.10289/) (2020)
- [MART: Improving LLM Safety with Multi-round Automatic Red-Teaming](https://mlsanity.com/paper/2311.07689/) (2023)
- [Deep Learning for Text Style Transfer: A Survey](https://mlsanity.com/paper/2011.00416/) (2020)
- [Detoxifying Language Models Risks Marginalizing Minority Voices](https://mlsanity.com/paper/2104.06390/) (2021)
- [Handling Bias in Toxic Speech Detection: A Survey](https://mlsanity.com/paper/2202.00126/) (2022)
- [Large Language Model Safety: A Holistic Survey](https://mlsanity.com/paper/2412.17686/) (2024)
- [From Languages to Geographies: Towards Evaluating Cultural Bias in Hate Speech Datasets](https://mlsanity.com/paper/2404.17874/) (2024)
- [Style-Specific Neurons for Steering LLMs in Text Style Transfer](https://mlsanity.com/paper/2410.00593/) (2024)
- [Towards A Friendly Online Community: An Unsupervised Style Transfer Framework for Profanity Redaction](https://mlsanity.com/paper/2011.00403/) (2020)
- [MultiParaDetox: Extending Text Detoxification with Parallel Data to New Languages](https://mlsanity.com/paper/2404.02037/) (2024)
- [Is LLM an Overconfident Judge? Unveiling the Capabilities of LLMs in Detecting Offensive Language with Annotation Disagreement](https://mlsanity.com/paper/2502.06207/) (2025)
- [DetoxLLM: A Framework for Detoxification with Explanations](https://mlsanity.com/paper/2402.15951/) (2024)
- [DefVerify: Do Hate Speech Models Reflect Their Dataset's Definition?](https://mlsanity.com/paper/2410.15911/) (2024)
- [K-HATERS: A Hate Speech Detection Corpus in Korean with Target-Specific Ratings](https://mlsanity.com/paper/2310.15439/) (2023)
- [Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation](https://mlsanity.com/paper/2412.14050/) (2024)
- [Trustworthy AI: Safety, Bias, and Privacy -- A Survey](https://mlsanity.com/paper/2502.10450/) (2025)
- [Unified Game Moderation: Soft-Prompting and LLM-Assisted Label Transfer for Resource-Efficient Toxicity Detection](https://mlsanity.com/paper/2506.06347/) (2025)
- [CoDAE: Adapting Large Language Models for Education via Chain-of-Thought Data Augmentation](https://mlsanity.com/paper/2508.08386/) (2025)
- [LLM in the Loop: Creating the ParaDeHate Dataset for Hate Speech Detoxification](https://mlsanity.com/paper/2506.01484/) (2025)
- [Beyond Over-Refusal: Scenario-Based Diagnostics and Post-Hoc Mitigation for Exaggerated Refusals in LLMs](https://mlsanity.com/paper/2510.08158/) (2025)
- [Text Detoxification: Data Efficiency, Semantic Preservation and Model Generalization](https://mlsanity.com/paper/2507.01050/) (2025)

## Built on by (most cited first, 1 of 1)

- [Who Pays More for Safety? Measuring the Disparate Cost of Safety Alignment across Languages](https://mlsanity.com/paper/2608.22490/) (2026)

## Abstract (arXiv)

While large language models (LLMs) have increasingly been applied to hate speech detoxification, the prompts often trigger safety alerts, causing LLMs to refuse the task. In this study, we systematically investigate false refusal behavior in hate speech detoxification and analyze the contextual and linguistic biases that trigger such refusals. We evaluate nine LLMs on both English and multilingual datasets, our results show that LLMs disproportionately refuse inputs with higher semantic toxicity and those targeting specific groups, particularly nationality, religion, and political ideology. Although multilingual datasets exhibit lower overall false refusal rates than English datasets, models still display systematic, language-dependent biases toward certain targets. Based on these findings, we propose a simple cross-translation strategy, translating English hate speech into Chinese for detoxification and back, which substantially reduces false refusals while preserving the original content, providing an effective and lightweight mitigation approach.