paper-with-me

홈 › Papers

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

2026-05-13 · Mokshit Surana, Archit Rathod, Akshaj Satishkumar 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.

📄 PDF Abstract BibTeX arXiv:2605.14087

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

FrenchToxicityPrompts: a Large Benchmark for Evaluating and Mitigating Toxicity in French Texts

2024-06-25 · Caroline Brun, Vassilina Nikoulina

Large language models (LLMs) are increasingly popular but are also prone to generating bias, toxic or harmful language, which can have detrimental effects on individuals and communities. Although most efforts is put to a…

TrustGPT: A Benchmark for Trustworthy and Responsible Large Language Models

2023-06-20 · Yue Huang, Qihui Zhang, Philip S. Y, Lichao Sun

Large Language Models (LLMs) such as ChatGPT, have gained significant attention due to their impressive natural language processing capabilities. It is crucial to prioritize human-centered principles when utilizing these…

Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation

2024-12-18 · Vera Neplenbroek, Arianna Bisazza, Raquel Fernández

Recent generative large language models (LLMs) show remarkable performance in non-English languages, but when prompted in those languages they tend to express higher harmful social biases and toxicity levels. Prior work …

Cross-Lingual TransferText Generation

Goodtriever: Adaptive Toxicity Mitigation with Retrieval-augmented Models

2023-10-11 · Luiza Pozzobon, Beyza Ermis, Patrick Lewis, Sara Hooker

Considerable effort has been dedicated to mitigating toxicity, but existing methods often require drastic modifications to model parameters or the use of computationally intensive auxiliary models. Furthermore, previous …

RetrievalText Generation

Reward Modeling for Mitigating Toxicity in Transformer-based Language Models

2022-02-19 · Farshid Faal, Ketra Schmitt, Jia Yuan Yu

Transformer-based language models are able to generate fluent text and be efficiently adapted across various natural language generation tasks. However, language models that are pretrained on large unlabeled web text cor…

Language ModelingLanguage ModellingText Generation