@inproceedings{chen-etal-2026-ban,
title = "Ban{\&}Pick: Enhancing Performance and Efficiency of {M}o{E}-{LLM}s via Smarter Routing",
author = "Chen, Yuanteng and
Wang, Peisong and
Shao, Yuantian and
Zeng, Nanxin and
Xu, Chang and
Cheng, Jian",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.678/",
pages = "14228--14253",
ISBN = "979-8-89176-462-0",
abstract = "Mixture-of-Experts (MoE) has become a key architecture for scaling large language models. Recent fine-grained MoE designs introduce hundreds of experts per layer, with multiple experts activated per token, enabling stronger specialization. However, during pre-training, routers are optimized mainly for stability and robustness: they converge prematurely and enforce balanced usage, limiting the full potential of model performance and efficiency at inference. In this work, we uncover two overlooked issues: (i) a few highly influential experts are underutilized due to premature and balanced routing decisions; and (ii) enforcing a fixed number of active experts per token introduces substantial redundancy. Instead of retraining or redesigning MoE architectures, we introduce Ban Pick, a training-free, plug-and-play strategy for smarter routing. Pick discovers and reinforces key experts, a small group with outsized impact on performance, leading to notable accuracy gains. Ban further dynamically prunes redundant experts based on layer and token sensitivity, delivering faster inference. Experiments on fine-grained MoE-LLMs (DeepSeek, Qwen3) across math, code, and general reasoning benchmarks demonstrate that Ban Pick delivers free performance gains and inference acceleration. For instance, on Qwen3-30B-A3B, it improves accuracy from 80.67 to 84.66 on AIME2024 and from 65.66 to 68.18 on GPQA-Diamond, while accelerating inference by 1.25{\texttimes} under the vLLM."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2026-ban">
<titleInfo>
<title>Ban&Pick: Enhancing Performance and Efficiency of MoE-LLMs via Smarter Routing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yuanteng</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peisong</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yuantian</namePart>
<namePart type="family">Shao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nanxin</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chang</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jian</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Mixture-of-Experts (MoE) has become a key architecture for scaling large language models. Recent fine-grained MoE designs introduce hundreds of experts per layer, with multiple experts activated per token, enabling stronger specialization. However, during pre-training, routers are optimized mainly for stability and robustness: they converge prematurely and enforce balanced usage, limiting the full potential of model performance and efficiency at inference. In this work, we uncover two overlooked issues: (i) a few highly influential experts are underutilized due to premature and balanced routing decisions; and (ii) enforcing a fixed number of active experts per token introduces substantial redundancy. Instead of retraining or redesigning MoE architectures, we introduce Ban Pick, a training-free, plug-and-play strategy for smarter routing. Pick discovers and reinforces key experts, a small group with outsized impact on performance, leading to notable accuracy gains. Ban further dynamically prunes redundant experts based on layer and token sensitivity, delivering faster inference. Experiments on fine-grained MoE-LLMs (DeepSeek, Qwen3) across math, code, and general reasoning benchmarks demonstrate that Ban Pick delivers free performance gains and inference acceleration. For instance, on Qwen3-30B-A3B, it improves accuracy from 80.67 to 84.66 on AIME2024 and from 65.66 to 68.18 on GPQA-Diamond, while accelerating inference by 1.25× under the vLLM.</abstract>
<identifier type="citekey">chen-etal-2026-ban</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.678/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>14228</start>
<end>14253</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Ban&Pick: Enhancing Performance and Efficiency of MoE-LLMs via Smarter Routing
%A Chen, Yuanteng
%A Wang, Peisong
%A Shao, Yuantian
%A Zeng, Nanxin
%A Xu, Chang
%A Cheng, Jian
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F chen-etal-2026-ban
%X Mixture-of-Experts (MoE) has become a key architecture for scaling large language models. Recent fine-grained MoE designs introduce hundreds of experts per layer, with multiple experts activated per token, enabling stronger specialization. However, during pre-training, routers are optimized mainly for stability and robustness: they converge prematurely and enforce balanced usage, limiting the full potential of model performance and efficiency at inference. In this work, we uncover two overlooked issues: (i) a few highly influential experts are underutilized due to premature and balanced routing decisions; and (ii) enforcing a fixed number of active experts per token introduces substantial redundancy. Instead of retraining or redesigning MoE architectures, we introduce Ban Pick, a training-free, plug-and-play strategy for smarter routing. Pick discovers and reinforces key experts, a small group with outsized impact on performance, leading to notable accuracy gains. Ban further dynamically prunes redundant experts based on layer and token sensitivity, delivering faster inference. Experiments on fine-grained MoE-LLMs (DeepSeek, Qwen3) across math, code, and general reasoning benchmarks demonstrate that Ban Pick delivers free performance gains and inference acceleration. For instance, on Qwen3-30B-A3B, it improves accuracy from 80.67 to 84.66 on AIME2024 and from 65.66 to 68.18 on GPQA-Diamond, while accelerating inference by 1.25× under the vLLM.
%U https://aclanthology.org/2026.findings-emnlp.678/
%P 14228-14253
Markdown (Informal)
[Ban&Pick: Enhancing Performance and Efficiency of MoE-LLMs via Smarter Routing](https://aclanthology.org/2026.findings-emnlp.678/) (Chen et al., Findings 2026)
ACL