多语言语言模型毒性检测与缓解策略综述
A Survey of Toxicity Detection and Mitigation Strategies for Multilingual Language Models
AI总结：
综述多语言大模型的毒性检测与缓解方法，包括威胁模型、检测方法（跨语言编码器、翻译流水线等）和缓解策略（数据过滤、偏好调优等），指出语言覆盖不均、文化依赖等挑战。
摘要：
大型语言模型（LLMs）越来越多地被部署到多种语言中，但其安全行为在不同语言和文化背景下仍不均衡。本综述综合了多语言LLMs的毒性检测与去毒化工作。我们首先分类了利用语言选择、翻译中介、代码切换、正字法变异、多轮交互和部署后微调来削弱安全对齐的威胁模型。然后，我们组织了任务形式（有毒到中性改写、毒性分类和有毒生成评估）、多语言检测方法（跨语言编码器、翻译流水线、表示级探测和基于LLM的检测器）以及缓解策略，涵盖数据过滤、监督和偏好调优、解码时引导、表示编辑和多语言护栏。在这些领域中，我们识别出持续的挑战：语言覆盖不均、伤害的文化依赖性定义、碎片化的评估协议，以及去毒化可能压制合法的方言或身份相关表达的风险。
英文摘要：
Large language models (LLMs) are increasingly deployed across languages, but their safety behavior remains uneven across linguistic and cultural contexts. This survey synthesizes work on toxicity detection and detoxification for multilingual LLMs. We first catalogue threat models that exploit language choice, translation pivots, code-switching, orthographic variation, multi-turn interaction, and post-deployment fine-tuning to weaken safety alignment. We then organize task formulations (toxic-to-neutral rewriting, toxicity classification, and toxic-generation evaluation), multilingual detection approaches (cross-lingual encoders, translation pipelines, representation-level probes, and LLM-based detectors), and mitigation strategies spanning data filtering, supervised and preference-based tuning, decoding-time steering, representation editing, and multilingual guardrails. Across these areas, we identify persistent challenges: uneven language coverage, culturally contingent definitions of harm, fragmented evaluation protocols, and the risk that detoxification suppresses legitimate dialectal or identity-related expression.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
