通过激活/关闭特定专家，无需微调即可控制大模型的安全与忠实性。
Steering MoE LLMs via Expert (De)Activation
- 通过对比正反样本的专家激活频率，定位行为相关专家。
- 安全性和忠实性最高提升27%，最差可降至-100%。
- 无需训练，适用于多数MoE模型，适合安全测试与风险评估。
大型语言模型中的混合专家（MoE）架构将每个标记路由至一组专用前馈网络（即专家）。本文提出SteerMoE框架，通过检测并控制与特定行为相关的专家来实现对MoE模型的引导。我们通过比较具有相反行为特征的输入对（如安全与不安全）中专家的激活频率，识别出关键行为专家。在推理阶段选择性地激活或禁用这些专家，即可控制模型的行为，如提高忠实性或安全性，而无需进行微调。在11个基准和6个主流LLM上的实验表明，该方法使安全性最高提升20%，忠实性最高提升27%。反之，恶意引导可使安全性下降41%，若结合现有越狱技术，安全性甚至降至-100%，完全绕过所有安全防护机制。总体而言，SteerMoE提供了一种轻量、高效且通用的运行时控制手段，同时揭示了MoE模型特有的安全隐患。
原文摘要 · Abstract (English)
Mixture-of-Experts (MoE) in Large Language Models (LLMs) routes each token through a subset of specialized Feed-Forward Networks (FFN), known as experts. We present SteerMoE, a framework to steer MoE models by detecting and controlling behavior-associated experts. We detect key experts by comparing how often they activate between paired inputs that demonstrate opposite behaviors (e.g., safe vs. unsafe). By selectively activating or deactivating such experts during inference, we control behaviors like faithfulness and safety without fine-tuning. Across 11 benchmarks and 6 LLMs, our steering raises safety by up to +20% and faithfulness by +27%. Alternatively, unsafe steering drops safety by -41% alone, and -100% when combined with existing jailbreak methods, bypassing all safety guardrails. Overall, SteerMoE offers a lightweight, effective, and widely applicable test-time control, while revealing unique vulnerabilities in MoE LLMs. https://github.com/adobe-research/SteerMoE
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。