The paper introduces Ban&Pick, a post‑training, plug‑and‑play routing strategy for Sparse Mixture‑of‑Experts large language models. It identifies and reinforces a small group of highly influential experts while dynamically pruning redundant ones, leading to accuracy gains across math, code, and reasoning benchmarks. Experiments on DeepSeek and Qwen3 show notable performance improvements and a 1.25× inference speedup without retraining or architectural changes.
By Yuanteng Chen, Peisong Wang, Yuantian Shao, Nanxin Zeng, Chang Xu, Jian Cheng
arXiv:2609.36222v1 Announce Type: new
Abstract: Large language models are increasingly expensive to serve. In large-scale serving systems, autoregressive decoding is often bottlenecked by transferrin...
By Ali Abbasi, Justin Shi, Soheil Kolouri
arXiv:2608. 10392v1 Announce Type: new Abstract: Mixture-of-experts (MoE) models have recently moved beyond routing a fixed number of complete experts.
By Gongli Zhang, Zhulin Liu, C. L. Philip Chen
arXiv:2608. 08853v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) routers commonly use the same scores both to select experts and to weight their already-computed outputs.
By Zongfei Li
arXiv:2607. 08780v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) models activate only a sparse subset of experts per token, yet consecutive tokens frequently activate different experts -- causing constant weight swapping between slow storage and fast memory on edge devices.
By Ali Kayyam
Mixture-of-Experts (MoE) models enable model scaling while maintaining low inference-time compute by activating only a subset of experts per token. However, conventional routing relies on a fixed top-k selection, forcing the model to spend the same compute regardless of how many experts are relevant.
DPA-MoE is a runtime precision control framework that allows Mixture-of-Experts (MoE) models to use any-precision routed experts. It encodes expert feed‑forward weights as prefix‑decodable chunks, profiles reconstruction residuals offline, and allocates chunks during inference based on live router gate scores. This enables a single encoded MoE checkpoint to operate at multiple effective weight budgets without retraining, improving KL divergence and downstream accuracy on Qwen3‑30B‑A3B and DeepSeekMoE‑16B with modest throughput overhead.
By Jaeyun Kim, Hoonki Lee, Susie Shin, Hyunmin Kim, Dahyun Lee, Sangyeob Kim
arXiv:2608.30158v1 Announce Type: cross
Abstract: Supervised fine-tuning (SFT) is the de facto standard for adapting large language models (LLMs) to target domains, but it often degrades the model's...
By Kwangmin Ki, Yunhun Nam, Jongheon Jeong, Jaehyung Kim
arXiv:2606. 10338v1 Announce Type: cross Abstract: Machine unlearning is increasingly important for large language models, yet unlearning in Mixture-of-Experts (MoE) architectures remains underexplored.
By Jingyi Xie, Yijun Lin, Yinjiang Xiong, Zhikun Zhang, Sai Li
arXiv:2608. 04401v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) models enable model scaling while maintaining low inference-time compute by activating only a subset of experts per token.
By Robin Pan, Raymond Liu, Daniel Fang, Adelina Andrei, Rosa Wu
Mixture-of-Experts (MoE) models reduce computation by routing each token to a small set of experts, but each selected expert is still used at a fixed precision. We present DPA-MoE, a runtime precision...
arXiv:2609.13058v1 Announce Type: new
Abstract: Reinforcement learning (RL) has become central to post-training of large language models. Recent advances in RL for Mixture-of-Experts (MoE) models hav...
By Hongyi He, Zhenghao Lin, Xiao Liu, Peng Cheng, Yan Lu, Yeyun Gong