MESA: Improving MoE Safety Alignment via Decentralized Expertise 事件

MESA: Improving MoE Safety Alignment via Decentralized Expertise arXiv:2606.00651v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) architectures scale Large Language Models (LLMs) efficiently, enabling greater capacity with reduced computational cost by dynamically routing inputs to relevant experts, yet introduce a critical vulnerability: Safety Sparsity, where safety capabilities concentrate in few experts, making them susceptible to adversarial bypassing. Meanwhile, conventional ali