Overview

Expert routing studies how models assign tokens, examples, or tasks to a small subset of experts, enabling conditional computation while storing many specialized networks. Common approaches include top-k token-choice and expert-choice gates, hierarchical and threshold-based routing, and adaptive policies driven by uncertainty, task information, or learned structure. Researchers test these methods in autoregressive language models, vision and multimodal systems, and diffusion models, measuring quality on benchmarks such as WikiText, MMLU, GSM8K, and ImageNet alongside load balance, latency, and memory use. A central tension is that balancing expert workloads can conflict with specialization and efficient execution, while analyses increasingly question whether routing patterns alone demonstrate genuine domain expertise or useful downstream effects.

Latest papers

CardsList

127 papers16 papers in the last four weeks, down 30% on the four weeks before. 0.1% of all new papers.

  1. Grouter: Decoupling Routing from Representation for Accelerated MoE Training

    Feb 22, 2026Yuqi Xu, Rizhen Hu, Zihan Liu +2Mixture-of-Experts ModelsExpert Routing

  2. Grassmannian Mixture-of-Experts: Concentration-Controlled Routing on Subspace Manifolds

    Feb 19, 2026Ibne Farabi Shihab, Sanjeda Akter, Anuj SharmaExpert RoutingMixture-of-Experts Language Models

  3. Phase-Aware Mixture of Experts for Agentic Reinforcement Learning

    Feb 19, 2026Shengtian Yang, Yu Li, Shuo He +4Reinforcement LearningRL for LLM Agents

  4. OmniMoE: An Efficient MoE by Orchestrating Atomic Experts at Scale

    Feb 5, 2026Jingze Shi, Zhangyang Peng, Yizhang Zhu +3Mixture-of-Experts InferenceLLM Inference Acceleration

  5. Routing-Aware Safety Alignment for Mixture-of-Experts Models

    Feb 4, 2026Jiacheng Liang, Yuhui Wang, Tanqiu Jiang +1Mixture-of-Experts Language ModelsLLM Alignment

  6. Stabilizing MoE Reinforcement Learning by Aligning Training and Inference Routers

    Oct 13, 2025Wenhan Ma, Hailin Zhang, Liang Zhao +4RL StabilityRL for Language Models