Mixture-of-Experts (MoE)

Researchers introduced the Groundhog Bit-Flip Attack (GBFA), the first bit-flip-based Denial-of-Wallet availability attack against Mixture-of-Experts (MoE) large language models, which inflates decodi…

Researchers introduced Task-Expert-Aware Supervision (TEXAS), a method that identifies task-relevant experts in Mixture-of-Experts (MoE) language models by comparing activations on successful versus f…