# Measuring and Mitigating Toxicity in Large Language Models: A Comprehensive Replication Study

arXiv: https://arxiv.org/abs/2605.14087 · HTML: https://mlsanity.com/paper/2605.14087/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 16 of its 26 references that are in the map.
- First posted on arXiv 13 May 2026.

## Builds on (most cited first, 16 of 16)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Constitutional AI: Harmlessness from AI Feedback](https://mlsanity.com/paper/2212.08073/) (2022)
- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [Plug and Play Language Models: A Simple Approach to Controlled Text Generation](https://mlsanity.com/paper/1912.02164/) (2019)
- [Bias and Fairness in Large Language Models: A Survey](https://mlsanity.com/paper/2309.00770/) (2023)
- [The Woman Worked as a Babysitter: On Biases in Language Generation](https://mlsanity.com/paper/1909.01326/) (2019)
- [JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models](https://mlsanity.com/paper/2404.01318/) (2024)
- [DExperts: Decoding-Time Controlled Text Generation with Experts and Anti-Experts](https://mlsanity.com/paper/2105.03023/) (2021)
- [FUDGE: Controlled Text Generation With Future Discriminators](https://mlsanity.com/paper/2104.05218/) (2021)
- [Challenges in Detoxifying Language Models](https://mlsanity.com/paper/2109.07445/) (2021)
- [Detoxifying Large Language Models via Knowledge Editing](https://mlsanity.com/paper/2403.14472/) (2024)
- [Unveiling the Implicit Toxicity in Large Language Models](https://mlsanity.com/paper/2311.17391/) (2023)
- [From One to Many: Expanding the Scope of Toxicity Mitigation in Language Models](https://mlsanity.com/paper/2403.03893/) (2024)
- [Realistic Evaluation of Toxicity in Large Language Models](https://mlsanity.com/paper/2405.10659/) (2024)
- [GTA: Gated Toxicity Avoidance for LM Performance Preservation](https://mlsanity.com/paper/2312.06122/) (2023)

## Abstract (arXiv)

Large Language Models (LLMs) trained on web-scale corpora inherently absorb toxic patterns from their training data. This leads to toxic degeneration where even innocuous prompts can trigger harmful outputs. This phenomenon poses significant risks for real-world deployments. Thus, necessitating effective mitigation strategies that should maintain model utility while ensuring safety. In this comprehensive replication study, we evaluate the efficacy of DExperts (Decoding-time Experts), which is an inference-time mitigation technique that steers generation without requiring model retraining. We structured our research into three systematic phases: (1) establishing baseline toxicity measurements using RealToxicityPrompts on standard GPT-2 models; then (2) implementing and evaluating DExperts to mitigate explicit toxicity; and finally (3) stress-testing the method against implicit hate speech using the adversarial ToxiGen dataset. Our empirical results confirm that while DExperts achieves near-perfect safety rates (100%) on explicit toxicity benchmarks, it exhibits brittleness against adversarial, implicit hate speech, with safety rates dropping to 98.5%. Furthermore, we quantify a critical trade-off. The method introduces a 10x latency penalty (from 0.2s to 2.0s per generation), posing challenges for real-time deployment scenarios. This study contributes to the growing body of work on AI safety by highlighting the robustness gap between explicit and implicit toxicity mitigation. We emphasize the need for more sophisticated approaches that generalize across diverse hate-speech patterns without incurring prohibitive computational costs.