Datasets:

fineset-io
/

mixture-of-experts-papers

id string	sources list	title string	abstract string	authors list	categories list	fields_of_study list	published_date timestamp[s]	url string	pdf_url string	arxiv_id string	doi string	citation_count int64	influential_citation_count int64	has_code bool	code_url string	venue string	quality_score float64
b2d211d97213b0fc16bb66c6b596f75a6106139ca5d4f2a89039bd82bee32db6	[ "arxiv" ]	SoftMoE: Soft Differentiable Routing for Mixture-of-Experts in LLMs	Sparse Mixture-of-Experts (MoE) architectures enable scaling LLM parameters under a fixed inference budget by activating only a small subset of experts via top-$k$ routing. While this preserves causality and suits autoregressive language models, the discrete top-$k$ operator is not differentiable, forcing a fixed numbe...	[ "Mikołaj Zasada", "Łukasz Struski", "Jacek Tabor", "Marcin Kurdziel" ]	[ "cs.LG", "cs.AI" ]	[]	2026-06-16T00:00:00	https://arxiv.org/abs/2606.17952	https://arxiv.org/pdf/2606.17952v1	2606.17952	null	0	0	false	null	null	0.35
09ea161a7d80f04e610defecd5c107e48b875ee666fd57935e11aea866e9c4a3	[ "arxiv", "semantic_scholar" ]	Tying the Loop -- Tied Expert Layers in Mixture-of-Experts Language Models	Mixture-of-Experts (MoE) architectures efficiently scale Large Language Models (LLMs) by activating only a small fraction of their experts per token, yet the full parameter count - dominated by the expert parameters - must be held in training and inference memory. To address this, we introduce Expert Tying, an architec...	[ "Martin Jaggi" ]	[ "cs.CL", "cs.AI", "cs.LG" ]	[ "Computer Science" ]	2026-06-15T00:00:00	https://arxiv.org/abs/2606.16825	https://arxiv.org/pdf/2606.16825v1	2606.16825	null	0	0	true	https://github.com/epfml/looped-moe	null	0.65
60430c52e8dd2ef3e39ce542fed4ea86469a732dcf5c9a28db75e4262c25810e	[ "arxiv" ]	MODE: Modality-Decomposed Expert-Level Mixed-Precision Quantization for MoE Multimodal LLMs	Mixture-of-Experts Multimodal Large Language Models (MoE-MLLMs) offer remarkable performance but incur prohibitive GPU memory costs, making compression essential. Among PTQ methods, expert-level mixed-precision quantization has proven effective for MoE-LLMs, yet suffers notable degradation on MoE-MLLMs due to two overl...	[ "Yuanteng Chen", "Peisong Wang", "Zhilei Liu", "Nanxin Zeng", "Yuantian Shao", "Shiqiang Lang", "Tao Liu", "Chuangyi Li", "Qinghao Hu", "Gang Li", "Jing Liu", "Jian Cheng" ]	[ "cs.LG", "cs.AI" ]	[]	2026-06-15T00:00:00	https://arxiv.org/abs/2606.17118	https://arxiv.org/pdf/2606.17118v1	2606.17118	null	0	0	false	null	null	0.35
0760d97bc48847ffd1e2eab76dfe12982f9d27e2aa4c0388b4ec18aa3a8f94ba	[ "arxiv", "semantic_scholar" ]	DynFS-MoE: Dynamic Functional-Structural Mixture-of-Experts for Post-Traumatic Epilepsy Diagnosis	Post-traumatic epilepsy (PTE) is a severe complication of traumatic brain injury (TBI), yet early identification remains challenging due to the complex structural and functional alterations it induces in the brain. To address this, we propose a dynamic multimodal Mixture-of-Experts (MoE) framework that integrates funct...	[ "Jun-En Ding", "Spencer Chen", "Henry Noren", "Daniel Valdivia", "Christine Yohn", "Suhina Patel", "Taylor Zink", "Hai Sun", "Feng Liu" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-06-15T00:00:00	https://arxiv.org/abs/2606.16203	https://arxiv.org/pdf/2606.16203v2	2606.16203	null	0	0	false	null	null	0.35
001bf195bfd8bc85e85ca1044ced24297d8a8a020106dcd182dd56bbfc3d1a8c	[ "arxiv", "semantic_scholar" ]	How to Score Experts for One-Shot MoE Expert Pruning: A Unified Formulation and Selection Principle	Mixture-of-Experts (MoE) language models reduce per-token computation through sparse expert activation, yet deployment still requires storing the full expert pool, making one-shot expert pruning a practical approach for reducing memory usage. Although effective, existing criteria are largely heuristic, and no single cr...	[ "Zongfang Liu", "Jinghui Zhang", "Zijian Ma", "Guangyi Chen", "Xin Yuan" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-06-14T00:00:00	https://arxiv.org/abs/2606.15716	https://arxiv.org/pdf/2606.15716v1	2606.15716	null	0	0	false	null	null	0.35
f7a30d41c0220f3e6c8c60ac612d4f5bc13f75c2419c990182ced4cf44089f4c	[ "arxiv", "semantic_scholar" ]	A Spatio-Temporal Expert Prefetching Framework for Efficient MoE-based LLM Inference	Mixture-of-Experts (MoE) based large language models (LLMs), such as Qwen and DeepSeek, have recently emerged as an effective approach to improving model capacity without proportionally increasing computational cost. By replacing the conventional feed-forward network in dense LLMs with a set of experts and activating o...	[ "Yingnan Zhao", "Razvan Bunescu", "Ahmed Louri", "Avinash Karanth", "Ke Wang" ]	[ "cs.AR", "cs.LG" ]	[ "Computer Science" ]	2026-06-13T00:00:00	https://arxiv.org/abs/2606.15453	https://arxiv.org/pdf/2606.15453v1	2606.15453	null	0	0	false	null	null	0.35
ef799393c45af71a9b784532b7b076d17c68b78a4b5d0df38e555133a08eff3e	[ "arxiv", "semantic_scholar" ]	A theoretical model for task routing in mixture-of-expert transformers	Mixture-of-experts (MoE) layers enable the scaling of transformer models while keeping the inference compute fixed. While task-expert specialization has been observed in empirical studies of frontier MoE transformer models, existing theoretical work analyzes this using continuous mixture models that cannot be used to m...	[ "Vinoth Nandakumar", "Yongli Xiang", "Yunzhi Yao", "Peike Li", "Tongliang Liu" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-06-12T00:00:00	https://arxiv.org/abs/2606.14398	https://arxiv.org/pdf/2606.14398v2	2606.14398	null	0	0	false	null	null	0.35
aeea47930489c3c29a15b5d032ebbd01137ca01018d6230a4d3ceabe6832ac05	[ "arxiv", "semantic_scholar" ]	Routing-Aware Expert Calibration for Machine Unlearning in Mixture-of-Experts Language Models	Machine unlearning is increasingly important for large language models, yet unlearning in Mixture-of-Experts (MoE) architectures remains underexplored. Unlike dense models, MoE architectures employ a router at each layer to assign each token to a sparse subset of experts. In this work, we observe that forget data often...	[ "Jingyi Xie", "Yijun Lin", "Yinjiang Xiong", "Zhikun Zhang", "Sai Li" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-06-09T00:00:00	https://arxiv.org/abs/2606.10338	https://arxiv.org/pdf/2606.10338v1	2606.10338	null	0	0	false	null	null	0.35
dba21527b12bb79fd47787f44818b7bea7a1972ff2267ec146dc4221afc9f563	[ "arxiv", "semantic_scholar" ]	STAR: Rethinking MoE Routing as Structure-Aware Subspace Learning	Mixture-of-Experts (MoE) scales model capacity efficiently by selectively routing inputs to a specialized subset of experts. However, input-expert specialization, the core motivation of MoE, critically depends on whether the router is actually aware of input structure. In practice, MoE routing is typically implemented ...	[ "Sumin Park", "Noseong Park" ]	[ "cs.AI", "cs.LG" ]	[ "Computer Science" ]	2026-06-07T00:00:00	https://arxiv.org/abs/2606.08814	https://arxiv.org/pdf/2606.08814v1	2606.08814	null	0	0	false	null	null	0.35
545471aea2dd74925794f33cf51f3913eeb80c642134aa3bb45f59ff1b2584ab	[ "arxiv", "semantic_scholar" ]	Reversible Foundations: Training a 120B Sparse MoE through State-Preserving Scaling	This paper reports on training a hundred-billion-parameter sparse mixture of experts on a single eight-GPU node, end to end. LightningLM 0.1V is a recurrence-backbone language model family grown in four stages from a small dense seed, through a 5B and a 9B mixture of experts, to a 120B model with 460 routed experts und...	[ "Rohan Shravan" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-06-05T00:00:00	https://arxiv.org/abs/2606.07404	https://arxiv.org/pdf/2606.07404v1	2606.07404	null	0	0	true	https://github.com/The-School-of-AI/LLM	null	0.65
5ea3182923a447700ff7de9a75dc9d912a92dbc16a5d4fdefccb458d42d9e167	[ "arxiv", "semantic_scholar" ]	YouZhi: Towards High-Concurrency Financial LLMs via Adaptive GQA-to-MLA Transition	Large language models (LLMs) drive significant financial innovations, yet their high-concurrency deployment is severely bottlenecked by KV cache memory overhead, which inflates infrastructure costs and throttles scalability. To address this, we propose YouZhi-LLM, a highly efficient financial LLM empowered by a compreh...	[ " PSBC LLM Team", " Huawei LLM Team", "Ruihan Long", "Junjie Wu", "Tianan Zhang", "Duo Zhang", "Yaozong Wu", "Jinbin Fu", "Chang Liu", "Zhentao Tang", "Wenshuang Yang", "Xin Wang", "Zhihao Song", "Ning Huang", "Wenjing Xu", "Shuai Zong", "Shupei Sun", "Sen Wang", "Jing Hu", "Bi...	[ "cs.CL" ]	[ "Computer Science" ]	2026-06-04T00:00:00	https://arxiv.org/abs/2606.05868	https://arxiv.org/pdf/2606.05868v1	2606.05868	null	0	0	false	null	null	0.35
51c0fafe4237deb7e37e634c4b1eb03507a834d97ea6de053d7c8a30353a889a	[ "arxiv", "semantic_scholar" ]	Value-and-Structure Alignment for Routing-Consistent Quantization of Mixture-of-Experts Models	Mixture-of-Experts (MoE) models scale foundation models efficiently by activating only a subset of experts for each token, but their large number of expert parameters still makes quantization essential for practical deployment. Unlike dense models, however, MoE models are sensitive to routing instability: small quantiz...	[ "Hancheol Park", "Geonho Lee", "Tairen Piao", "Tae-Ho Kim" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-06-04T00:00:00	https://arxiv.org/abs/2606.05688	https://arxiv.org/pdf/2606.05688v1	2606.05688	null	0	0	false	null	null	0.35
d98765a0cc7fec86538d4fb4aba9d678b35fe350ba967ff7df21501a168e1d55	[ "arxiv", "semantic_scholar" ]	Less is MoE: Trimming Experts in Domain-Specialist Language Models	Mixture-of-Experts (MoE) models achieve strong performance through conditional computation, but their large parameter footprint poses deployment challenges. Prior MoE compression approaches catastrophically fail when evaluated on general-purpose benchmarks beyond commonsense reasoning. We trace this failure to the gran...	[ "Haoze He", "Xinkai Zou", "Xuan Jiang", "Xingyuan Ding", "Ao Qu", "Juncheng Billy Li", "Heather Miller" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-06-04T00:00:00	https://arxiv.org/abs/2606.05538	https://arxiv.org/pdf/2606.05538v1	2606.05538	null	0	0	false	null	null	0.35
ff4d98f6b0b06cd072a826b0088f9a2be1ac76f474e701014d7d0e6caf8817ff	[ "arxiv", "semantic_scholar" ]	SHAPE: Coalition-Aware Expert Pruning for Sparse Mixture-of-Experts LLMs	Sparse Mixture-of-Experts (MoE) large language models achieve strong quality with low per-token compute, yet their deployment is often limited by the memory wall: the full expert pool must remain resident to support token-dependent routing. Expert pruning is a direct remedy, but prior criteria typically score experts i...	[ "Yuhao Zhang" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-06-03T00:00:00	https://arxiv.org/abs/2606.09886	https://arxiv.org/pdf/2606.09886v1	2606.09886	null	0	0	true	https://github.com/Alizen-1009/Shapley-Moe	null	0.65
07e219c4a0b5e8637c1a78a853167272aff18638a7b9f17ce412b5cfbeeeafa9	[ "arxiv", "semantic_scholar" ]	Expert-Aware Causal Tracing of Factual Recall in Sparse MoE Language Models	Causal tracing of factual recall has been studied predominantly in dense transformer language models, where interventions localize information flow to layers or feed-forward modules. Sparse mixture-of-experts (MoE) language models introduce a sharper question: when a factual prediction is mediated by a routed MoE block...	[ "Yuetian Lu", "Ali Modarressi", "Yihong Liu", "Hinrich Schütze" ]	[ "cs.CL", "cs.LG" ]	[ "Computer Science" ]	2026-06-02T00:00:00	https://arxiv.org/abs/2606.03780	https://arxiv.org/pdf/2606.03780v1	2606.03780	null	0	0	false	null	null	0.35
ffa7dd593613a170a8aefdf8639b244173ec134c9309d4131f852ba6e5d3818e	[ "arxiv", "semantic_scholar" ]	Sparse Mixture-of-Experts Reward Models Learn Interpretable and Specialized Experts for Personalized Preference Modeling	Preference modeling plays a central role in reinforcement learning from human feedback (RLHF), enabling large language models (LLMs) to align with human values. However, most existing approaches assume a universal reward function, neglecting the diversity and heterogeneity of human preferences. To address this limitati...	[ "Yifan Wang", "Jinyi Mu", "Mayank Jobanputra", "Yu Wang", "Ji-Ung Lee", "Soyoung Oh", "Isabel Valera", "Vera Demberg" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-06-02T00:00:00	https://arxiv.org/abs/2606.04284	https://arxiv.org/pdf/2606.04284v1	2606.04284	null	0	0	false	null	null	0.35
94fad39b46a45fc010ee9ac33de2fc46ee763585ba9ed646e199cbca786238f8	[ "arxiv", "semantic_scholar" ]	When Model Merging Breaks Routing: Training-Free Calibration for MoE	Model merging has emerged as a cost-effective approach for consolidating the capabilities of multiple LLMs without retraining. However, existing merging techniques, largely based on linear parameter arithmetic or optimization, struggle when applied to Mixture-of-Experts (MoE) architectures. We identify a critical failu...	[ "Canbin Huang", "Tianyuan Shi", "Xiaojun Quan", "Jingang Wang", "Jianfei Zhang", "Qifan Wang" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-06-02T00:00:00	https://arxiv.org/abs/2606.03391	https://arxiv.org/pdf/2606.03391v1	2606.03391	null	0	0	true	https://github.com/huangcb01/HARC	null	0.65
84a56ce3f3c4253499659b6510485f8b267ea5ca1eb3e8bd49c5ee207fb2c611	[ "arxiv", "semantic_scholar" ]	ProbMoE: Differentiable Probabilistic Routing for Mixture-of-Experts	Mixture-of-Experts (MoE) models scale by activating only a small subset of experts per token. However, training such models remains challenging because top-$k$ routing is discrete and non-differentiable, requiring gradient estimators for expert selection whose design remains a central open problem. We introduce ProbMoE...	[ "Heng Zhao", "Zilei Shao", "Guy Van den Broeck", "Zhe Zeng" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-06-01T00:00:00	https://arxiv.org/abs/2606.01509	https://arxiv.org/pdf/2606.01509v1	2606.01509	null	0	0	false	null	null	0.35
27ad4d392659179ee22e669923b194321238275ce5ff05d1c3f0ecddffe1e993	[ "arxiv", "semantic_scholar" ]	DAG-MoE: From Simple Mixture to Structural Aggregation in Mixture-of-Experts	Mixture-of-Experts (MoE) models have become a leading approach for decoupling parameter count from computational cost in large language models, yet effectively scaling MoE performance remains a challenge. Prior work shows that fine-grained experts enlarge the space of expert combinations and improve flexibility, but th...	[ "Jiarui Feng", "Hanqing Zeng", "Karish Grover", "Ruizhong Qiu", "Yinglong Xia", "Qiang Zhang", "Qifan Wang", "Ren Chen", "Dongqi Fu", "Jiayi Liu", "Zhoukai Zhao", "Xiangjun Fan", "Benyu Zhang", "Yixin Chen" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-05-31T00:00:00	https://arxiv.org/abs/2606.01062	https://arxiv.org/pdf/2606.01062v1	2606.01062	null	1	0	false	null	null	0.35
ea8b3767e1b2bfc8d1f5ccdccda8c68c9dfc5ae03e3f5ed4cb22d65f47ee6fa2	[ "arxiv", "semantic_scholar" ]	PR2: Predictive Routing Replay for MoE-Based LLM Reinforcement Learning	Mixture of Experts (MoE) Large Language Models (LLMs) achieve strong performance at scale. However, reinforcement learning (RL) on MoE-based LLMs often suffers from training instability. A root cause is router drift, i.e., expert activations can change drastically across model updates and differ between disaggregated r...	[ "Daize Dong", "Junlin Chen", "Haolong Jia", "Jiang Liu", "Jiawei Wu", "Huanwei Di", "Jialian Wu", "Zhengzhong Liu", "Zicheng Liu", "Emad Barsoum", "Dimitris N. Metaxas", "Hongyi Wang" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-05-29T00:00:00	https://arxiv.org/abs/2606.00395	https://arxiv.org/pdf/2606.00395v2	2606.00395	null	0	0	false	null	null	0.35
a8d3312b5c7c7d574e91e4e921cbaea3ebbcfb099210631933c18bf8e7f457ab	[ "arxiv", "semantic_scholar" ]	Understanding Safety-Sensitive Expert Behavior in Mixture-of-Experts LLMs	Mixture-of-Experts (MoE) LLMs rely on sparse, router-driven expert activation, yet how safety alignment interacts with routed expert specialization remains underexplored. A common intuition is that safety behavior may be controlled by routing harmful requests to distinct refusal-oriented experts. In this work, we provi...	[ "Zhibo Zhang", "Yuxi Li", "Zhen Ouyang", "Ling Shi", "Kailong Wang" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-05-28T00:00:00	https://arxiv.org/abs/2605.29708	https://arxiv.org/pdf/2605.29708v1	2605.29708	null	0	0	false	null	null	0.35
a5831e669d725d1e4b2635777049fc108b3f0888bc1894c190c828f40ed697b8	[ "arxiv", "semantic_scholar" ]	Leveraging Routing Dynamics in Mixture-of-Experts Models for Efficient Language Adaptation	Mixture-of-Experts (MoE) models are widely used to scale language models, yet their expert routing behavior and adaptation in a multilingual setting remain underexplored. In this work, we study multilingual routing dynamics during continual pre-training of an English-centric MoE model on a multilingual corpus, analyzin...	[ "Aditi Khandelwal", "Marius Mosbach", "Verna Dankers", "Siva Reddy", "Golnoosh Farnadi" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-05-28T00:00:00	https://arxiv.org/abs/2605.29714	https://arxiv.org/pdf/2605.29714v1	2605.29714	null	0	0	true	https://github.com/aditi184/moe-routing-adaptation	null	0.65
ea437a6b57a0a04601aa92b74f7282904035a423378aebc18bb6ea2e6025f335	[ "arxiv", "semantic_scholar" ]	ConMoE: Expert-Pool Consolidation via Prototype Reassignment for MoE Compression	Mixture-of-Experts (MoE) language models reduce per-token computation but still require storing and serving all experts, making deployment memory-intensive. Existing post-training compression methods mainly shrink this cost by pruning experts or merging their weights. We formulate post-training MoE compression as exper...	[ "Yilun Yao", "Jiaming Pan", "Elsie Dai", "Peizhuang Cong", "Yaoming Li", "Tong Yang" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-05-28T00:00:00	https://arxiv.org/abs/2605.29350	https://arxiv.org/pdf/2605.29350v1	2605.29350	null	0	0	false	null	null	0.35
bb990fba3c9f75cb10fb9cff57b6e223123b7e0bec5772da42cf4b9e791470bd	[ "arxiv", "semantic_scholar" ]	Routing-Aligned Fine-Tuning for Multilingual Downstream Tasks in Mixture-of-Experts Models	Mixture-of-Experts (MoE) models have emerged as a dominant paradigm for efficient LLM scaling, yet adapting them to non-English downstream tasks remains challenging. Existing fine-tuning approaches treat MoE models as monolithic learners, ignoring the heterogeneous routing structure that develops during pretraining. We...	[ "Guanzhi Deng", "Kuan Wu", "Haibo Wang", "Shing Yin Wong", "Sichun Luo", "Linqi Song" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-05-27T00:00:00	https://arxiv.org/abs/2605.28306	https://arxiv.org/pdf/2605.28306v1	2605.28306	null	0	0	false	null	null	0.35
b932e82ed1d79b177eee5497205ae1d57a047b5427c0a0d9220ebbea8ca2a6e5	[ "arxiv", "semantic_scholar" ]	FPMoE: A Sparse Mixture-of-Experts Approach to Functional Code Generation	Despite rapid progress in LLM-based code generation, existing models are predominantly trained on imperative languages, leaving functional programming languages (FPLs) such as Haskell, OCaml, and Scala chronically underexplored, with even frontier models performing substantially worse on FPLs. Fine-tuning is a natural ...	[ "Loc Pham", "Lang Hong Nguyet Anh", "Thanh Le-Cong" ]	[ "cs.PL", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-05-27T00:00:00	https://arxiv.org/abs/2605.27849	https://arxiv.org/pdf/2605.27849v1	2605.27849	null	0	0	true	null	null	0.65
94f6014df324797f71eeeefb71f1626b7bdc1e2ad7a7b57ffe80fb74d93f3baf	[ "arxiv", "semantic_scholar" ]	ReMoE: Boosting Expert Reuse through Router Fine-Tuning in Memory-Constrained MoE LLM Inference	Fine-grained Mixture-of-Experts (MoE) models sparsely activate only a subset of experts per token, reducing activated computation while maintaining high model capacity. However, in memory-constrained inference scenarios, only a small set of experts can be cached. Experts not in the cache must be fetched from slow exter...	[ "Xiongwei Zhu", "Xiaojian Liao", "Tianyang Jiang", "Yusen Zhang", "Liang Wang", "Limin Xiao" ]	[ "cs.LG", "cs.AI", "cs.DC" ]	[ "Computer Science" ]	2026-05-26T00:00:00	https://arxiv.org/abs/2605.27081	https://arxiv.org/pdf/2605.27081v1	2605.27081	null	0	0	true	https://github.com/BUAA-OSCAR/ReMoE	null	0.65
8922eb922492a3eaa9cfc80b9951558467d90f2d451467f8cded4d4aa4e68a59	[ "arxiv", "semantic_scholar" ]	BioFact-MoE: Biologically Factorized Mixture of Experts for Vision-Language Prognostic Modeling in Hepatocellular Carcinoma	Hepatocellular carcinoma (HCC) is biologically heterogeneous, shaped by the interplay between hepatic functional reserve and tumor-related oncologic factors; thus, similar survival outcomes may reflect fundamentally different underlying biological processes. Prognostic modeling in HCC is informed by rich multimodal inf...	[ "Junlin Yang", "Tian Yu", "Nicha C. Dvornek", "Yuexi Du", "Peiyu Duan", "Annabella Shewarega", "Lawrence H. Staib", "James S. Duncan", "Julius Chapiro" ]	[ "cs.CV", "cs.AI", "cs.LG" ]	[ "Computer Science" ]	2026-05-25T00:00:00	https://arxiv.org/abs/2605.26376	https://arxiv.org/pdf/2605.26376v1	2605.26376	null	0	0	true	https://github.com/jy-639/BioFact-MoE	null	0.65
01766860289bf7462da1adc8c4096eca2fb38bf947e91d55fa8beee387e6a605	[ "arxiv", "semantic_scholar" ]	RotMoLE: Enhancing Mixture of Low-Rank Experts through Rotational Gating Mechanism	While Large Language Models (LLMs) are commonly fine-tuned to handle domain-specific tasks before being applied to vertical applications, adapting them to complex scenarios with diverse specialized knowledge remains challenging. Meanwhile, Mixture-of-Experts (MoE) architecture has risen as a crucial paradigm for traini...	[ "Mengyang Sun", "Maochuan Dou", "Tao Feng", "Dan Zhang", "Yihao Wang", "Junpeng Liu", "Yifan Zhu", "Jie Tang" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-25T00:00:00	https://arxiv.org/abs/2605.25565	https://arxiv.org/pdf/2605.25565v1	2605.25565	null	0	0	false	null	null	0.35
40184c69a06ed460ef6db4dc81c9c355c502f5d8f843034a42073362e46abad2	[ "arxiv", "semantic_scholar" ]	RouteScan: A Non-Intrusive Approach to Auditing MoE LLMs Safety via Expert Routing Telemetry	Mixture-of-Experts (MoE) architectures have become an increasingly important paradigm for scaling Large Language Models (LLMs). As MoE models are increasingly deployed in real-world services, safety auditing becomes necessary to verify whether these models produce or facilitate harmful behaviors during operation. Howev...	[ "Bo Lv", "Zhiheng Xu", "KeDong Xiu", "Ruyi Ding", "Tianhang Zheng", "Zhibo Wang", "Kui Ren" ]	[ "cs.CR", "cs.AR", "cs.CL", "cs.LG" ]	[ "Computer Science" ]	2026-05-24T00:00:00	https://arxiv.org/abs/2605.24817	https://arxiv.org/pdf/2605.24817v1	2605.24817	null	0	0	true	null	null	0.65
e4b2536b0fe4f721202b1ac28b188887b92fcec60eec13ffeac945e8035a6138	[ "arxiv", "semantic_scholar" ]	AME-TS: Anchored Mixture-of-Experts for Time Series Forecasting	Time series forecasting models are increasingly scaled through large Transformer backbones, yet most existing approaches process all series through a shared dense computation path despite substantial heterogeneity in temporal structure. Mixture-of-Experts (MoE) offers a natural alternative by enabling conditional compu...	[ "Rui Wang", "Renhao Xue", "Ray Razi", "Huan Song", "Hannah R. Marlowe" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-05-24T00:00:00	https://arxiv.org/abs/2605.25166	https://arxiv.org/pdf/2605.25166v1	2605.25166	null	0	0	false	null	null	0.35
9fb1f5cdf7e5fa9789b3b962a72a70295676685c01d248a432e1fe4e1b426229	[ "arxiv", "semantic_scholar" ]	Safety-Oriented Routing Analysis of Mixtral MoE Under Benign and Harmful Prompts	Sparse mixture-of-experts (MoE) language models activate only a small subset of parameters for each token, making router behavior a central part of model computation. This paper studies routing behavior of Mixtral 8x7B-Instruct under benign and harmful prompts using two complementary signals: activation-based routing s...	[ "Md Nurul Absar Siddiky" ]	[ "cs.AI", "cs.CR" ]	[ "Computer Science" ]	2026-05-22T00:00:00	https://arxiv.org/abs/2605.24270	https://arxiv.org/pdf/2605.24270v1	2605.24270	null	0	0	false	null	null	0.35
5c1fa6f5cd729e4de740c84a6528488c846d8af0e5aa39e6696bd1094bc2b8e6	[ "arxiv", "semantic_scholar" ]	GEMQ: Global Expert-Level Mixed-Precision Quantization for MoE LLMs	Mixture-of-Experts Large Language Models (MoE-LLMs) achieve strong performance but incur substantial memory overhead due to massive expert parameters. Mixed-precision quantization mitigates this cost by allocating expert-wise bit-widths based on their importance, approaching the accuracy-memory Pareto frontier and enab...	[ "Jianing Deng", "Song Wang", "Dongwei Wang", "Zijie Liu", "Tianlong Chen", "Huanrui Yang", "Jingtong Hu" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-21T00:00:00	https://arxiv.org/abs/2605.23078	https://arxiv.org/pdf/2605.23078v1	2605.23078	null	0	0	true	https://github.com/jndeng/GEMQ	null	0.65
4e5862621c61c300f9ab95b4f8fa94ef1ce14225651c8649ef5568fb11cbfc53	[ "arxiv", "semantic_scholar" ]	Beyond Routing: Characterising Expert Tuning and Representation in Vision Mixture-of-Experts	Mixture-of-Experts (MoE) models are often interpreted by analysing which categories are routed to which experts. However, routing alone does not reveal what each expert actually encodes. We train sparsely-gated convolutional MoE models with a contrastive objective on natural images and characterise expert specialisatio...	[ "Gene Tangtartharakul", "Katherine R. Storrs" ]	[ "cs.CV", "cs.AI" ]	[ "Computer Science" ]	2026-05-20T00:00:00	https://arxiv.org/abs/2605.20610	https://arxiv.org/pdf/2605.20610v1	2605.20610	null	0	0	false	null	null	0.35
c08824c4c5c184721917a400dc5a9eb4e992dd4df7274f6748aad169dd56983a	[ "arxiv", "semantic_scholar" ]	Task-Routed Mixture-of-Experts with Cognitive Appraisal for Implicit Sentiment Analysis	Implicit sentiment analysis is challenging because sentiment toward an aspect is often inferred from events rather than expressed through explicit opinion words. Existing models typically learn from the final polarity label, which provides limited guidance for reasoning about sentiment from the context. Motivated by co...	[ "Yaping Chai", "Haoran Xie", "Joe S. Qin" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-05-20T00:00:00	https://arxiv.org/abs/2605.20916	https://arxiv.org/pdf/2605.20916v1	2605.20916	null	0	0	true	https://github.com/yaping166/TRMoE-ISA	null	0.65
227f742ce30306c31364321e7f9b321d3fc52f5124666dbfaa0e69a91e0c4219	[ "arxiv", "semantic_scholar" ]	GEM: GPU-Variability-Aware Expert to GPU Mapping for MoE Systems	Mixture-of-Expert (MoE) models enable efficient inference by employing smaller experts and activating only a subset of them per token. MoE serving engines distribute experts across multiple GPUs and route tokens to appropriate GPUs at inference time based on experts activated. They process tokens in lock-step fashion, ...	[ "Sourish Wawdhane", "Avinash Kumar", "Poulami Das" ]	[ "cs.DC", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-05-19T00:00:00	https://arxiv.org/abs/2605.19945	https://arxiv.org/pdf/2605.19945v1	2605.19945	null	0	0	false	null	null	0.35
dad487bf361cd50b76dc26f154f08f386fcd2fa8687ecb069b8b7c00adc80424	[ "arxiv", "semantic_scholar" ]	CP-MoE: Consistency-Preserving Mixture-of-Experts for Continual Learning	Catastrophic forgetting remains a major obstacle to continual learning in large language models (LLMs) and vision--language models (VLMs). Although Mixture-of-Experts (MoE) architectures offer an efficient path to scaling, existing LoRA-based MoE continual learning methods still face a fundamental trade-off: they eithe...	[ "Yang Liu", "Toan Nguyen", "Flora D. Salim" ]	[ "cs.LG", "cs.AI", "cs.CL", "cs.CV" ]	[ "Computer Science" ]	2026-05-18T00:00:00	https://arxiv.org/abs/2605.20247	https://arxiv.org/pdf/2605.20247v1	2605.20247	null	0	0	false	null	null	0.35
70007cbe46cb60004f951353a022582aa2407d27647592c0e1a2e8f864162cde	[ "arxiv", "semantic_scholar" ]	Post-Trained MoE Can Skip Half Experts via Self-Distillation	Mixture-of-Experts (MoE) scales language models efficiently through sparse expert activation, and its dynamic variant further reduces computation by adjusting the activated experts in an input-dependent manner. Existing dynamic MoE methods usually rely on pre-training from scratch or task-specific adaptation, leaving t...	[ "Xingtai Lv", "Li Sheng", "Kaiyan Zhang", "Yichen You", "Siyan Gao", "Xueheng Luo", "Yuxin Zuo", "Yuchen Fan", "Junlin Yang", "Ganqu Cui", "Bingning Wang", "Fan Yang", "Youbang Sun", "Ning Ding", "Bowen Zhou" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-05-18T00:00:00	https://arxiv.org/abs/2605.18643	https://arxiv.org/pdf/2605.18643v2	2605.18643	null	0	0	false	null	null	0.35
da51de550d416c9c388c92e7cd00169479fad6d651d3fc9185ff0029a63770e9	[ "arxiv", "semantic_scholar" ]	CoX-MoE: Coalesced Expert Execution for High-Throughput MoE Inference with AMX-Enabled CPU-GPU Co-Execution	The Mixture-of-Experts (MoE) architecture improves computational efficiency via sparse expert activation, but throughput-oriented inference faces substantial GPU memory pressure due to a significant parameter size and intermediate data. Prior works attempt to mitigate this using expert offloading with micro-batching or...	[ "Muyoung Son", "Yi Chen", "Seungjae Yoo", "Soongyu Choi", "Joo-Young Kim" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-05-18T00:00:00	https://arxiv.org/abs/2605.17889	https://arxiv.org/pdf/2605.17889v2	2605.17889	10.1145/3770743.3804296	0	0	false	null	null	0.35
4c8039055e72df82136c1b13514d5cf1bfe2e310acd1ae152938e49722960cf7	[ "arxiv", "semantic_scholar" ]	Stable Routing for Mixture-of-Experts in Class-Incremental Learning	Class-incremental learning (CIL) requires models to learn new classes sequentially while preserving prior knowledge. Recently, approaches that combine pre-trained models with mixture-of-experts (MoE) have received increasing attention in CIL: they typically expand experts during learning and employ a router to assign w...	[ "Zirui Guo", "Quan Cheng", "Da-Wei Zhou", "Lijun Zhang" ]	[ "cs.CV", "cs.LG" ]	[ "Computer Science" ]	2026-05-17T00:00:00	https://arxiv.org/abs/2605.17571	https://arxiv.org/pdf/2605.17571v1	2605.17571	null	0	0	false	null	null	0.35
1a128d3435cc8c20b2102a526e0fdd9f27106e523f2de09de19370cc4d21c2e5	[ "arxiv", "semantic_scholar" ]	Mixture of Experts for Low-Resource LLMs	Mixture-of-Experts (MoE) architectures enable efficient model scaling, yet expert routing behavior across underrepresented languages remains poorly understood. We analyze routing dynamics in two architecturally distinct MoE models -- a pure Transformer (Qwen3-30B-A3B) and a hybrid Mamba-Transformer (Nemotron-3-Nano-30B...	[ "Ori Bar Joseph", "Smadar Arvatz", "Noam Kayzer", "Dan Revital", "Sarel Weinberger" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-05-17T00:00:00	https://arxiv.org/abs/2605.17598	https://arxiv.org/pdf/2605.17598v1	2605.17598	null	0	0	false	null	null	0.35
c4fa328f487b6388fb835929bfcd9985a0e1eb046dc49085a0df37beba1a0925	[ "arxiv", "semantic_scholar" ]	When Does Sparse MoE Help in Vision? The Role of Backbone Compute Leverage in Sparse Routing	Mixture-of-Experts (MoE) networks promise favorable accuracy-compute trade-offs, yet practical vision deployments are hindered by expert collapse and limited end-to-end efficiency gains. We study when sparse top-$k$ routing with hard capacity constraints helps in vision classification, evaluated under multi-seed protoc...	[ "Libo Sun", "Po-wei Harn", "Peixiong He", "Xiao Qin" ]	[ "cs.CV", "cs.LG" ]	[ "Computer Science" ]	2026-05-15T00:00:00	https://arxiv.org/abs/2605.15484	https://arxiv.org/pdf/2605.15484v1	2605.15484	null	0	0	true	https://github.com/libophd/sparse-moe-vision-rho	null	0.65
13cbd8f25fffdf1d761bad64c2d409be957da44ae36702503e6f85f3ccef5ccd	[ "arxiv", "semantic_scholar" ]	Scalable Knowledge Editing for Mixture-of-Experts LLMs via Tensor-Structured Updates	Knowledge editing (KE) provides a lightweight alternative to repeated fine-tuning of LLMs. However, most existing KE methods target dense feed-forward layers, while modern LLMs increasingly adopt Mixture-of-Experts (MoE) architectures for their superior memory footprint and inference efficiency. This mismatch leaves a ...	[ "Roman Maksimov", "Vladimir Aletov", "Dmitry Bylinkin", "Daniil Medyakov", "Vladimir Solodkin", "Aleksandr Beznosikov" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-05-15T00:00:00	https://arxiv.org/abs/2605.16686	https://arxiv.org/pdf/2605.16686v1	2605.16686	null	0	0	false	null	null	0.35
ff481e8d16ed8916ca1a6a7f9c311f5a43c80c1a8d1d876f8d51d0256bd657da	[ "arxiv", "semantic_scholar" ]	BEAM: Binary Expert Activation Masking for Dynamic Routing in MoE	Mixture-of-Experts (MoE) architectures enhance the efficiency of large language models by activating only a subset of experts per token. However, standard MoE employs a fixed Top-K routing strategy, leading to redundant computation and suboptimal inference latency. Existing acceleration methods either require costly re...	[ "Juntong Wu", "Jialiang Cheng", "Qishen Yin", "Yue Dai", "Yuliang Yan", "Fuyu Lv", "Ou Dan", "Li Yuan" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-05-14T00:00:00	https://arxiv.org/abs/2605.14438	https://arxiv.org/pdf/2605.14438v1	2605.14438	null	0	0	false	null	null	0.35
22e44138f60f689ab135f54fedccf041a78538f1ce546105b0d75c3ad5d18b48	[ "arxiv", "semantic_scholar" ]	RQ-MoE: Residual Quantization via Mixture of Experts for Efficient Input-Dependent Vector Compression	Vector quantization is a fundamental tool for compressing high-dimensional embeddings, yet existing multi-codebook methods rely on static codebooks that limit expressiveness under heterogeneous data geometry. While recent dynamic quantizers like QINCo adapt codebooks to individual inputs and improve expressiveness, the...	[ "Zhengjia Zhong", "Shuyan Ke", "Zaizhou Lin", "Jiaqi Song", "Hongyi Lan", "Hui Li" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-05-14T00:00:00	https://arxiv.org/abs/2605.14359	https://arxiv.org/pdf/2605.14359v1	2605.14359	null	0	0	true	https://github.com/KDEGroup/RQ-MoE	null	0.65
fb668949e35e901f640695146e3ed2504eb19bacc7ffc3005cf0a06d42017960	[ "arxiv", "semantic_scholar" ]	Eradicating Negative Transfer in Multi-Physics Foundation Models via Sparse Mixture-of-Experts Routing	Scaling Scientific Machine Learning (SciML) toward universal foundation models is bottlenecked by negative transfer: the simultaneous co-training of disparate partial differential equation (PDE) regimes can induce gradient conflict, unstable optimization, and plasticity loss in dense neural operators. In particular, br...	[ "Ellwil Sharma", "Arastu Sharma" ]	[ "cs.LG", "cs.AI", "physics.comp-ph" ]	[ "Computer Science", "Physics" ]	2026-05-14T00:00:00	https://arxiv.org/abs/2605.15179	https://arxiv.org/pdf/2605.15179v1	2605.15179	null	0	0	false	null	null	0.35
a6770ec8df8337956d27e296bb66489720f2535c356bab623db4fdbb4bbf49b8	[ "arxiv", "semantic_scholar" ]	HodgeCover: Higher-Order Topological Coverage Drives Compression of Sparse Mixture-of-Experts	Sparse Mixture-of-Experts (MoE) layers route tokens through a handful of experts, and learning-free compression of these layers reduces inference cost without retraining. A subtle obstruction blocks every existing compressor in this family: three experts can each be pairwise compatible yet form an irreducible cycle whe...	[ "Tao Zhong", "Dongzhe Zheng", "Christine Allen-Blanchette" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-05-13T00:00:00	https://arxiv.org/abs/2605.13997	https://arxiv.org/pdf/2605.13997v1	2605.13997	null	1	0	false	null	null	0.35
ca9d22c67454bb9e3ad117448f5d725bb9a7f6b5e76e6fe6753af11d9c73b748	[ "arxiv", "semantic_scholar" ]	ROMER: Expert Replacement and Router Calibration for Robust MoE LLMs on Analog Compute-in-Memory Systems	Large language models (LLMs) with mixture-of-experts (MoE) architectures achieve remarkable scalability by sparsely activating a subset of experts per token, yet their frequent expert switching creates memory bandwidth bottlenecks that compute-in-memory (CIM) architectures are well-suited to mitigate. However, analog C...	[ "Wenyong Zhou", "Yuannuo Feng", "Yizhe Chen", "Taiqiang Wu", "Wendong Xu", "Wenbo Qi", "Zhengwu Liu", "Wang Kang", "Ngai Wong" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-12T00:00:00	https://arxiv.org/abs/2605.11800	https://arxiv.org/pdf/2605.11800v1	2605.11800	null	0	0	false	null	null	0.35
2649365b9ef754a8eded64ee29fef529cb948f7fa086ca1c8b4872c31eeb11f0	[ "arxiv", "semantic_scholar" ]	Sparse Mixture-of-Experts Routing in Visual Diffusion Transformers:Diagnosis, Boundary Calibration and Evolutionary Roadmap from Routing Collapse to Selective Deadlock	This paper systematically diagnoses the training failure modes of Token-Choice sparse Mixture-of-Experts (MoE) on video Diffusion Transformers. Starting from a pretrained dense model of about 5 billion parameters, we convert it into an MoE architecture following three laws: routed experts exactly clone the original FFN...	[ "Haiying Sha" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-05-12T00:00:00	https://arxiv.org/abs/2605.19378	https://arxiv.org/pdf/2605.19378v1	2605.19378	null	0	0	false	null	null	0.35
6b1c74f8d098790e24800236e07e6dc62c4f7a2c3c649473e66cc15b77ffdb04	[ "arxiv", "semantic_scholar" ]	Fast MoE Inference via Predictive Prefetching and Expert Replication	The Mixture of Experts (MoE) architecture has become a fundamental building block in state-of-the-art large language models (LLMs), improving domain-specific expertise in LLMs and scaling model capacity without proportionally increasing their computational overhead. However, MoE inference often suffers from suboptimal ...	[ "Ankit Jyothish", "Ali Jannesari", "Aishwarya Sarkar", "Joseph Zuber" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-05-12T00:00:00	https://arxiv.org/abs/2605.11537	https://arxiv.org/pdf/2605.11537v1	2605.11537	null	0	0	false	null	null	0.35
1fad0cb93aaa669ab230021fef265de93672f8b0a0afdef8d6262fe91d8051cf	[ "arxiv", "semantic_scholar" ]	Routers Learn the Geometry of Their Experts: Geometric Coupling in Sparse Mixture-of-Experts	Sparse Mixture-of-Experts (SMoE) models enable scaling language models efficiently, but training them remains challenging, as routing can collapse onto few experts and auxiliary load-balancing losses can reduce specialization. Motivated by these hurdles, we study how routing decisions in SMoEs are formed mechanisticall...	[ "Sagi Ahrac", "Noya Hochwald", "Mor Geva" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-12T00:00:00	https://arxiv.org/abs/2605.12476	https://arxiv.org/pdf/2605.12476v1	2605.12476	null	0	0	false	null	null	0.35
8d9b1dd13670e546e863d6e0f40204db7e5c9633e3b2671484dc53a2537d4059	[ "arxiv", "semantic_scholar" ]	DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices	While Mixture-of-Experts (MoE) scales model capacity without proportionally increasing computation, its massive total parameter footprint creates significant storage and memory-access bottlenecks, which hinder efficient end-side deployment that simultaneously requires high performance, low computational cost, and small...	[ "Chenyang Song", "Weilin Zhao", "Xu Han", "Chaojun Xiao", "Yingfa Chen", "Zhiyuan Liu" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-11T00:00:00	https://arxiv.org/abs/2605.10933	https://arxiv.org/pdf/2605.10933v3	2605.10933	null	0	0	true	https://github.com/thunlp/DECO	null	0.65
d10b208fcadce025522f523023fd498a10c0a51a05cdf48242fd8efd07cd05ed	[ "arxiv", "semantic_scholar" ]	ReLibra: Routing-Replay-Guided Load Balancing for MoE Training in Reinforcement Learning	Load imbalance is a long-standing challenge in Mixture-of-Experts (MoE) training and is exacerbated in reinforcement learning (RL) for LLMs, where hot experts can shift frequently across micro-batches. Existing MoE training systems rely on historical loads to predict future expert demand, making them less effective und...	[ "Chao Jin", "Xinming Wei", "Yinmin Zhong", "Chengxu Yang", "Bingyang Wu", "Ruidong Zhu", "Zili Zhang", "Yuliang Liu", "Xin Jin" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-05-09T00:00:00	https://arxiv.org/abs/2605.08639	https://arxiv.org/pdf/2605.08639v1	2605.08639	null	1	0	false	null	null	0.35
1fc1067bcee9a2f92d682f5284e06b1aa6237f8e1d6cb66e4ca222724a6c1627	[ "arxiv", "semantic_scholar" ]	SDG-MoE: Signed Debate Graph Mixture-of-Experts	Sparse MoE models achieve a good balance between capacity and compute by routing each token to a small subset of experts. However, in most MoE architectures, once a token is routed, the selected experts process it independently and their outputs are combined via a weighted sum. This leaves open whether enabling communi...	[ "Stepan Kulibaba", "Kirill Labzin", "Artem Dzhalilov", "Roman Pakhomov", "Oleg Svidchenko", "Alexander Gasnikov", "Aleksei Shpilman" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-05-08T00:00:00	https://arxiv.org/abs/2605.08322	https://arxiv.org/pdf/2605.08322v2	2605.08322	null	0	0	false	null	null	0.35
a633d22994699e0be8dab405d558f163fdd5c9b9ab4f2785c337994bdc6b87f1	[ "arxiv", "semantic_scholar" ]	When Are Experts Misrouted? Counterfactual Routing Analysis in Mixture-of-Experts Language Models	Mixture-of-Experts (MoE) language models route each token to a small subset of experts, but whether the routes selected by a trained top-$k$ router are good ones is rarely evaluated directly. Holding the model fixed, we compare each standard route against sampled equal-compute alternatives for the same token and score ...	[ "Youngsik Yoon", "Siwei Wang", "Wei Chen", "Jungseul Ok" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-05-08T00:00:00	https://arxiv.org/abs/2605.07260	https://arxiv.org/pdf/2605.07260v1	2605.07260	null	0	0	false	null	null	0.35
f481399e160c099ead5bb8154483676e307c4fcd79045bc7c46274fbda829724	[ "arxiv", "semantic_scholar" ]	Hierarchical Mixture-of-Experts with Two-Stage Optimization	Sparse Mixture-of-Experts (MoE) models scale capacity by routing each token to a small subset of experts. However, their routers exhibit a fundamental trade-off: strong load balancing can suppress expert specialization, while aggressive diversity often causes routing collapse. We propose Hi-MoE, a grouped MoE framework...	[ "Gleb Molodtsov", "Alexander Miasnikov", "Aleksandr Beznosikov" ]	[ "cs.LG", "cs.AI", "math.OC" ]	[ "Computer Science", "Mathematics" ]	2026-05-08T00:00:00	https://arxiv.org/abs/2605.08292	https://arxiv.org/pdf/2605.08292v1	2605.08292	null	2	0	false	null	null	0.35
1a97c99a67d652e5ee5a3f234f74d12fe527edd4454a4e3b997aa540a99c406f	[ "arxiv", "semantic_scholar" ]	Expert Routing for Communication-Efficient MoE via Finite Expert Banks	Resource-efficient machine learning increasingly uses sparse Mixture-of-Experts (MoE) architectures, where the gate acts as both a learning component and a routing interface controlling computation, communication, and accuracy. Motivated by finite-rate interpretations of MoE gating, we treat the gate as a stochastic ch...	[ "Mohammad Reza Deylam Salehi", "Ali Khalesi" ]	[ "cs.LG", "cs.IT" ]	[ "Computer Science", "Mathematics" ]	2026-05-06T00:00:00	https://arxiv.org/abs/2605.05278	https://arxiv.org/pdf/2605.05278v1	2605.05278	null	1	0	false	null	null	0.35
9a4cde77ec3dd676758cdd0d50b62bbb05d28df120647d91081b91c81773855a	[ "arxiv", "semantic_scholar" ]	GEM: Graph-Enhanced Mixture-of-Experts with ReAct Agents for Dialogue State Tracking	Dialogue State Tracking (DST) requires precise extraction of structured information from multi-domain conversations, a task where Large Language Models (LLMs) struggle despite their impressive general capabilities. We present GEM (Graph-Enhanced Mixture-of-Experts), a novel framework that combines language models and g...	[ "Ziqi Zhu", "Adithya Suresh", "Tomal Deb", "Iman Abbasnejad" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-05-06T00:00:00	https://arxiv.org/abs/2605.04449	https://arxiv.org/pdf/2605.04449v1	2605.04449	null	0	0	false	null	null	0.35
8430e6b515734c28e56ee4aa7f5017caeff983fc15780479b33a60c4974f63ff	[ "arxiv", "semantic_scholar" ]	Adaptive Inverted-Index Routing for Granular Mixtures-of-Experts	Mixture-of-experts (MoE) models enable scalable transformer architectures by activating only a subset of experts per token. Recent evidence suggests that performance improves with increasingly granular experts, i.e., many small experts instead of a few large ones. However, this regime substantially increases routing co...	[ "Klaus-Rudolf Kladny", "Maximilian Mordig", "Bernhard Schölkopf", "Michael Muehlebach" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-05-06T00:00:00	https://arxiv.org/abs/2605.04952	https://arxiv.org/pdf/2605.04952v1	2605.04952	null	0	0	false	null	null	0.35
4a789bbc361b9a23dc2c111f7aab7771a2a8b3dfd6d497c7cd89572a24c2d920	[ "arxiv", "semantic_scholar" ]	Misrouter: Exploiting Routing Mechanisms for Input-Only Attacks on Mixture-of-Experts LLMs	Mixture-of-Experts (MoE) architectures have emerged as a leading paradigm for scaling large language models through sparse, routing-based computation. However, this design introduces a new attack surface: the routing mechanism that determines which experts process each input. Prior work shows that manipulating routing ...	[ "Zekun Fei", "Zihao Wang", "Weijie Liu", "Ruiqi He", "Jianing Geng", "Zheli Liu", "XiaoFeng Wang" ]	[ "cs.CR" ]	[ "Computer Science" ]	2026-05-06T00:00:00	https://arxiv.org/abs/2605.04446	https://arxiv.org/pdf/2605.04446v1	2605.04446	null	1	0	true	null	null	0.65
d0fbb94fe1b345c19327ec4cb7bf5e19fec57a13b879ce35a83e0b053420dd30	[ "arxiv", "semantic_scholar" ]	Soft-to-Hard Routing in Sparse Mixture-of-Experts Models	Softmax routing approaches hard top-1 routing as the temperature tends to zero, but the limiting passage is singular at router ties. This paper develops a boundary-layer calculus for this soft-to-hard limit in population squared-loss mixture-of-experts regression. For a router with logits $a_k(x;φ)$, the relevant local...	[ "Reza Rastegar" ]	[ "cs.LG", "cs.AI", "math.PR" ]	[ "Computer Science", "Mathematics" ]	2026-05-04T00:00:00	https://arxiv.org/abs/2605.02124	https://arxiv.org/pdf/2605.02124v2	2605.02124	null	1	0	false	null	null	0.35
3f90697cba1df1cf4315203cd881232aa35e1d68d1cea139ba7729df7b039269	[ "arxiv", "semantic_scholar" ]	RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs	Safety alignment is critical for the responsible deployment of large language models (LLMs). As Mixture-of-Experts (MoE) architectures are increasingly adopted to scale model capacity, understanding their safety robustness becomes essential. Existing adversarial attacks, however, have notable limitations. Prompt-based ...	[ "Zhiyuan Xu", "Joseph Gardiner", "Sana Belguith", "Lichao Wu" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-05-01T00:00:00	https://arxiv.org/abs/2605.02946	https://arxiv.org/pdf/2605.02946v1	2605.02946	null	2	0	false	null	null	0.35
8c1937e90e48523a24100793ad084d7331784fc48de9bb09da6a08281f384202	[ "arxiv", "semantic_scholar" ]	Marco-MoE: Open Multilingual Mixture-of-Expert Language Models with Efficient Upcycling	We present Marco-MoE, a suite of fully open multilingual sparse Mixture-of-Experts (MoE) models. Marco-MoE features a highly sparse design in which only around 5\% of the total parameters are activated per input token. This extreme sparsity, combined with upcycling from dense models, enables efficient pre-training on 5...	[ "Fan Jiang", "Yu Zhao", "Chenyang Lyu", "Tianqi Shi", "Yichao Du", "Feihu Jiang", "Longyue Wang", "Weihua Luo" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-04-28T00:00:00	https://arxiv.org/abs/2604.25578	https://arxiv.org/pdf/2604.25578v1	2604.25578	10.48550/arXiv.2604.25578	0	0	false	null	arXiv.org	0.55
2bcd8c7bb424f1f5ac37005fa451537977c5e42d962cf5d6a4cf577909ace47f	[ "arxiv", "semantic_scholar" ]	SMoES: Soft Modality-Guided Expert Specialization in MoE-VLMs	Mixture-of-Experts (MoE) has become a prevalent backbone for large vision-language models (VLMs), yet how modality-specific signals should guide expert routing remains under-explored. Existing routing strategies are either hand-crafted or modality-agnostic, relying on idealized priors that ignore the layer-dependent mo...	[ "Zi-Hao Bo", "Yaqian Li", "Anzhou Hou", "Rinyoichi Takezoe", "Ertao Zhao", "Tianxiang Pan", "Jiale Yan", "Mo Guang", "Kaiwen Long" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-04-27T00:00:00	https://arxiv.org/abs/2604.23996	https://arxiv.org/pdf/2604.23996v1	2604.23996	10.48550/arXiv.2604.23996	0	0	false	null	arXiv.org	0.55
819d2eb3f0cc7702fb53daebe933e4d2df11701be6e3d8b852cd64f6db70677c	[ "arxiv", "semantic_scholar" ]	Mixture of Heterogeneous Grouped Experts for Language Modeling	Large Language Models (LLMs) based on Mixture-of-Experts (MoE) are pivotal in industrial applications for their ability to scale performance efficiently. However, standard MoEs enforce uniform expert sizes,creating a rigidity that fails to align computational costs with varying token-level complexity. While heterogeneo...	[ "Zhicheng Ma", "Xiang Liu", "Zhaoxiang Liu", "Ning Wang", "Yi Shen", "Kai Wang", "Shuming Shi", "Shiguo Lian" ]	[ "cs.CL", "cs.AI", "cs.LG" ]	[ "Computer Science" ]	2026-04-25T00:00:00	https://arxiv.org/abs/2604.23108	https://arxiv.org/pdf/2604.23108v2	2604.23108	10.48550/arXiv.2604.23108	0	0	true	https://github.com/UnicomAI/MoHGE	arXiv.org	0.85
566f4f5f44350f4e0e6a5acac83ecd1af628a006d254a6ee196e364f2df5bbc8	[ "arxiv", "semantic_scholar" ]	Teacher-Guided Routing for Sparse Vision Mixture-of-Experts	Recent progress in deep learning has been driven by increasingly large-scale models, but the resulting computational cost has become a critical bottleneck. Sparse Mixture of Experts (MoE) offers an effective solution by activating only a small subset of experts for each input, achieving high scalability without sacrifi...	[ "Masahiro Kada", "Ryota Yoshihashi", "Satoshi Ikehata", "Rei Kawakami", "Ikuro Sato" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-04-23T00:00:00	https://arxiv.org/abs/2604.21330	https://arxiv.org/pdf/2604.21330v1	2604.21330	10.48550/arXiv.2604.21330	0	0	false	null	arXiv.org	0.55
55caf3cf72eeaa5ce78d52e1fbae8cb056bc4095a00fa07f0c9a64fb97f87395	[ "arxiv", "semantic_scholar" ]	Rethinking Cross-Domain Evaluation for Face Forgery Detection with Semantic Fine-grained Alignment and Mixture-of-Experts	Nowadays, visual data forgery detection plays an increasingly important role in social and economic security with the rapid development of generative models. Existing face forgery detectors still can't achieve satisfactory performance because of poor generalization ability across datasets. The key factor that led to th...	[ "Yuhan Luo", "Tao Chen", "Decheng Liu" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-04-23T00:00:00	https://arxiv.org/abs/2604.21478	https://arxiv.org/pdf/2604.21478v1	2604.21478	10.48550/arXiv.2604.21478	0	0	true	https://github.com/Yuhan-Luo/Semantic-Fine-grained-Alignment-and-Mixture-of-Experts	arXiv.org	0.85
4bf5ad019c0d6007cf73a483f2544581c4a7d6e6b9926707837a1b09660fe97b	[ "arxiv", "semantic_scholar" ]	Expert Upcycling: Shifting the Compute-Efficient Frontier of Mixture-of-Experts	Mixture-of-Experts (MoE) has become the dominant architecture for scaling large language models: frontier models routinely decouple total parameters from per-token computation through sparse expert routing. Scaling laws show that under fixed active computation, model quality scales predictably with total parameters, an...	[ "Chaitanya Dwivedi", "Binxuan Huang", "Himanshu Gupta", "Pratik Jayarao", "Neeraj Varshney", "Bing Yin" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-04-21T00:00:00	https://arxiv.org/abs/2604.19835	https://arxiv.org/pdf/2604.19835v2	2604.19835	10.48550/arXiv.2604.19835	0	0	false	null	arXiv.org	0.55
01530af61de47c1c6ddb2d6dd4cf2f476c0ea5088b48a9cb811976e63ee8151d	[ "arxiv", "semantic_scholar" ]	Polysemantic Experts, Monosemantic Paths: Routing as Control in MoEs	An LLM's residual stream is both state and instruction: it encodes the current context and determines the next transformation. We introduce a parameter-free decomposition for Mixture-of-Experts models that splits each layer's hidden state into a control signal that causally drives routing and an orthogonal content chan...	[ "Charles Ye", "Bo Yuan", "Lee Sharkey" ]	[ "cs.AI", "cs.CL", "cs.LG" ]	[ "Computer Science" ]	2026-04-20T00:00:00	https://arxiv.org/abs/2604.17837	https://arxiv.org/pdf/2604.17837v1	2604.17837	10.48550/arXiv.2604.17837	0	0	false	null	arXiv.org	0.55
d97b7228a5e6972e30bf0371705bce4431a378635a81fcd6177d9e358b73dc45	[ "arxiv", "semantic_scholar" ]	MoE-nD: Per-Layer Mixture-of-Experts Routing for Multi-Axis KV Cache Compression	KV cache memory is the dominant bottleneck for long-context LLM inference. Existing compression methods each act on a single axis of the four-dimensional KV tensor -- token eviction (sequence), quantization (precision), low-rank projection (head dimension), or cross-layer sharing -- but apply the same recipe to every l...	[ "Libo Sun", "Peixiong He", "Po-Wei Harn", "Xiao Qin" ]	[ "cs.LG", "cs.CL" ]	[ "Computer Science" ]	2026-04-20T00:00:00	https://arxiv.org/abs/2604.17695	https://arxiv.org/pdf/2604.17695v1	2604.17695	10.48550/arXiv.2604.17695	0	0	false	null	arXiv.org	0.55
d954a955521897820e5aa9035a3c5ef113ed0907200869317db0fbad6bbf5c28	[ "arxiv", "semantic_scholar" ]	Efficient Mixture-of-Experts LLM Inference with Apple Silicon NPUs	Apple Neural Engine (ANE) is a dedicated neural processing unit (NPU) present in every Apple Silicon chip. Mixture-of-Experts (MoE) LLMs improve inference efficiency via sparse activation but are challenging for NPUs in three ways: expert routing is unpredictable and introduces dynamic tensor shapes that conflict with ...	[ "Afsara Benazir", "Felix Xiaozhu Lin" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-04-20T00:00:00	https://arxiv.org/abs/2604.18788	https://arxiv.org/pdf/2604.18788v1	2604.18788	10.48550/arXiv.2604.18788	0	0	false	null	arXiv.org	0.55
1f15bb1f449f9f066a6756ebc68206b172abc53e40840f9528eb595eb9f71c86	[ "arxiv", "semantic_scholar" ]	CoGR-MoE: Concept-Guided Expert Routing with Consistent Selection and Flexible Reasoning for Visual Question Answering	Visual Question Answering (VQA) requires models to identify the correct answer options based on both visual and textual evidence. Recent Mixture-of-Experts (MoE) methods improve option reasoning by grouping similar concepts or routing based on examples. However, unstable routing can lead to inconsistent expert selectio...	[ "Xiyin Zeng", "Yi Lu", "Hao Wang" ]	[ "cs.CV", "cs.AI" ]	[ "Computer Science" ]	2026-04-18T00:00:00	https://arxiv.org/abs/2604.16930	https://arxiv.org/pdf/2604.16930v1	2604.16930	10.48550/arXiv.2604.16930	0	0	false	null	arXiv.org	0.5477
c12ba085126bb19f4fab5642a6d2bede59b56436b172294c944991f85470b666	[ "arxiv", "semantic_scholar" ]	Geometric Routing Enables Causal Expert Control in Mixture of Experts	Sparse Mixture-of-Experts (MoE) models scale parameters while fixing active computation per token, but the specialization of individual experts remains opaque. In a companion paper we showed that routing topology is quality-neutral: five structurally different configurations converge to statistically equivalent languag...	[ "Ivan Ternovtsii", "Yurii Bilak" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-04-15T00:00:00	https://arxiv.org/abs/2604.14434	https://arxiv.org/pdf/2604.14434v1	2604.14434	10.48550/arXiv.2604.14434	2	0	false	null	arXiv.org	0.5443
5f3fc99c6e87648ba63eb1ea1054f6a70cb1a86a5fe5b10668d54916d3aed793	[ "arxiv", "semantic_scholar" ]	Design and Behavior of Sparse Mixture-of-Experts Layers in CNN-based Semantic Segmentation	Sparse mixture-of-experts (MoE) layers have been shown to substantially increase model capacity without a proportional increase in computational cost and are widely used in transformer architectures, where they typically replace feed-forward network blocks. In contrast, integrating sparse MoE layers into convolutional ...	[ "Svetlana Pavlitska", "Haixi Fan", "Konstantin Ditschuneit", "J. Marius Zöllner" ]	[ "cs.CV", "cs.LG" ]	[ "Computer Science" ]	2026-04-15T00:00:00	https://arxiv.org/abs/2604.13761	https://arxiv.org/pdf/2604.13761v1	2604.13761	10.48550/arXiv.2604.13761	1	0	true	https://github.com/KASTEL-MobilityLab/moe-layers/	arXiv.org	0.8411
b0219f75b9bef0ac1fd423254e916728187e7f45f7bcd7b76eb1ceb2b6692b32	[ "arxiv", "semantic_scholar" ]	Equifinality in Mixture of Experts: Routing Topology Does Not Determine Language Modeling Quality	Sparse Mixture-of-Experts (MoE) architectures employ increasingly sophisticated routing mechanisms -- learned routers, multi-hop trajectories, token-dependent gating. We ask: does routing topology actually determine language modeling quality? We build a geometric MoE (ST-MoE) using cosine-similarity routing against lea...	[ "Ivan Ternovtsii", "Yurii Bilak" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-04-15T00:00:00	https://arxiv.org/abs/2604.14419	https://arxiv.org/pdf/2604.14419v1	2604.14419	10.48550/arXiv.2604.14419	2	0	false	null	arXiv.org	0.5443
f0f54cf02d286315396a5f7841e83c8918ad678054028476027e8ef005312bb7	[ "arxiv", "semantic_scholar" ]	Awakening Dormant Experts:Counterfactual Routing to Mitigate MoE Hallucinations	Sparse Mixture-of-Experts (MoE) models have achieved remarkable scalability, yet they remain vulnerable to hallucinations, particularly when processing long-tail knowledge. We identify that this fragility stems from static Top-$k$ routing: routers tend to favor high-frequency patterns over rare factual associations. Co...	[ "Wentao Hu", "Yanbo Zhai", "Xiaohui Hu", "Mingkuan Zhao", "Shanhong yu", "Xue Liu", "Kaidong Yu", "Shuangyong Song", "Xuelong Li" ]	[ "cs.LG", "cs.AI" ]	[ "Computer Science" ]	2026-04-15T00:00:00	https://arxiv.org/abs/2604.14246	https://arxiv.org/pdf/2604.14246v2	2604.14246	10.48550/arXiv.2604.14246	0	0	false	null	arXiv.org	0.5443
cc887166da6f4fabc95e4d1944081bca5adde00f1ddbeaff02947254849188cb	[ "arxiv", "semantic_scholar" ]	Nucleus-Image: Sparse MoE for Image Generation	We present Nucleus-Image, a text-to-image generation model that establishes a new Pareto frontier in quality-versus-efficiency by matching or exceeding leading models on GenEval, DPG-Bench, and OneIG-Bench while activating only approximately 2B parameters per forward pass. Nucleus-Image employs a sparse mixture-of-expe...	[ "Chandan Akiti", "Ajay Modukuri", "Murali Nandan Nagarapu", "Gunavardhan Akiti", "Haozhe Liu" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-04-14T00:00:00	https://arxiv.org/abs/2604.12163	https://arxiv.org/pdf/2604.12163v1	2604.12163	10.48550/arXiv.2604.12163	2	0	true	null	arXiv.org	0.8394
19216a298011a671051d513fc9249a307f483ffe3ddfa83e6cd2b125a13ac9b9	[ "arxiv", "semantic_scholar" ]	The Myth of Expert Specialization in MoEs: Why Routing Reflects Geometry, Not Necessarily Domain Expertise	Mixture of Experts (MoEs) are now ubiquitous in large language models, yet the mechanisms behind their "expert specialization" remain poorly understood. We show that, since MoE routers are linear maps, hidden state similarity is both necessary and sufficient to explain expert usage similarity, and specialization is the...	[ "Xi Wang", "Soufiane Hayou", "Eric Nalisnick" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-04-10T00:00:00	https://arxiv.org/abs/2604.09780	https://arxiv.org/pdf/2604.09780v1	2604.09780	10.48550/arXiv.2604.09780	4	1	false	null	arXiv.org	0.5385
219ab6374d4916c7e505d4a442b1d9194b034d9b45dbe0e21202df45f44a28bc	[ "arxiv", "semantic_scholar" ]	Seeing but Not Thinking: Routing Distraction in Multimodal Mixture-of-Experts	Multimodal Mixture-of-Experts (MoE) models have achieved remarkable performance on vision-language tasks. However, we identify a puzzling phenomenon termed Seeing but Not Thinking: models accurately perceive image content yet fail in subsequent reasoning, while correctly solving identical problems presented as pure tex...	[ "Haolei Xu", "Haiwen Hong", "Hongxing Li", "Rui Zhou", "Yang Zhang", "Longtao Huang", "Hui Xue", "Yongliang Shen", "Weiming Lu", "Yueting Zhuang" ]	[ "cs.CV", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-04-09T00:00:00	https://arxiv.org/abs/2604.08541	https://arxiv.org/pdf/2604.08541v1	2604.08541	10.48550/arXiv.2604.08541	2	0	false	null	arXiv.org	0.5374
bcaca74006d18d962c840c438ad987027885ca125da1bca37a25c44c29b0b3f1	[ "arxiv", "semantic_scholar" ]	Alloc-MoE: Budget-Aware Expert Activation Allocation for Efficient Mixture-of-Experts Inference	Mixture-of-Experts (MoE) has become a dominant architecture for scaling large language models due to their sparse activation mechanism. However, the substantial number of expert activations creates a critical latency bottleneck during inference, especially in resource-constrained deployment scenarios. Existing approach...	[ "Baihui Liu", "Kaiyuan Tian", "Wei Wang", "Zhaoning Zhang", "Linbo Qiao", "Dongsheng Li" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-04-09T00:00:00	https://arxiv.org/abs/2604.08133	https://arxiv.org/pdf/2604.08133v1	2604.08133	10.48550/arXiv.2604.08133	0	0	false	null	arXiv.org	0.5374
760850ffc8ebc3ec9dcf72c5f256543f82110a0e11e15cfd6ec875c25db04962	[ "arxiv", "semantic_scholar" ]	MoE Routing Testbed: Studying Expert Specialization and Routing Behavior at Small Scale	Sparse Mixture-of-Experts (MoE) architectures are increasingly popular for frontier large language models (LLM) but they introduce training challenges due to routing complexity. Fully leveraging parameters of an MoE model requires all experts to be well-trained and to specialize in non-redundant ways. Assessing this, h...	[ "Tobias Falke", "Nicolas Anastassacos", "Samson Tan", "Chankrisna Richy Meas", "Chandana Satya Prakash", "Nitesh Sekhar", "M Saiful Bari", "Krishna Kompella", "Gamaleldin F. Elsayed" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-04-08T00:00:00	https://arxiv.org/abs/2604.07030	https://arxiv.org/pdf/2604.07030v1	2604.07030	10.48550/arXiv.2604.07030	0	0	false	null	arXiv.org	0.5363
17a3008c95aaa5ef05bc9849d9b88037f9e3cdc5efb3c45430990406554f06fa	[ "arxiv", "semantic_scholar" ]	Region-Graph Optimal Transport Routing for Mixture-of-Experts Whole-Slide Image Classification	Multiple Instance Learning (MIL) is the dominant framework for gigapixel whole-slide image (WSI) classification in computational pathology. However, current MIL aggregators route all instances through a shared pathway, constraining their capacity to specialise across the pathological heterogeneity inherent in each slid...	[ "Xin Tian", "Jiuliu Lu", "Ephraim Tsalik", "Bart Wanders", "Colleen Knoth", "Julian Knight" ]	[ "cs.CV", "cs.AI", "eess.IV" ]	[ "Computer Science", "Engineering" ]	2026-04-08T00:00:00	https://arxiv.org/abs/2604.07298	https://arxiv.org/pdf/2604.07298v1	2604.07298	10.48550/arXiv.2604.07298	0	0	false	null	arXiv.org	0.5363
6cbfcfa948788f6c102de75255ea891114756250d835c043a13f3614ea31947e	[ "arxiv", "semantic_scholar" ]	Does a Global Perspective Help Prune Sparse MoEs Elegantly?	Empirical scaling laws for language models have encouraged the development of ever-larger LLMs, despite their growing computational and memory costs. Sparse Mixture-of-Experts (MoEs) offer a promising alternative by activating only a subset of experts per forward pass, improving efficiency without sacrificing performan...	[ "Zeliang Zhang", "Nikhil Ghosh", "Jiani Liu", "Bin Yu", "Xiaodong Liu" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-04-08T00:00:00	https://arxiv.org/abs/2604.06542	https://arxiv.org/pdf/2604.06542v1	2604.06542	10.48550/arXiv.2604.06542	0	0	false	null	arXiv.org	0.5363
37b3ae67c869171ae3fedaa6cd198cf44c35a23518bbf8c04a0566fb99b8f2b4	[ "arxiv", "semantic_scholar" ]	Cross-Platform Fused MoE Dispatch in Triton: Portable Expert Routing Without CUDA	Mixture-of-Experts (MoE) architectures power the majority of frontier large language models, but their inference is bottlenecked by irregular memory access patterns and expert routing overhead. Existing optimized MoE kernels (Megablocks, Tutel, FasterMoE) are implemented in CUDA and locked to NVIDIA hardware. We presen...	[ "Subhadip Mitra" ]	[ "cs.DC", "cs.ET", "cs.PF" ]	[ "Computer Science" ]	2026-04-07T00:00:00	https://arxiv.org/abs/2605.23911	https://arxiv.org/pdf/2605.23911v1	2605.23911	null	0	0	true	https://github.com/bassrehab/triton-kernels	null	0.6324
0e7e5d481f91939e921cd9aafb0daf1676d464af31136a40495e8e91cafdc7a4	[ "arxiv", "semantic_scholar" ]	Do Domain-specific Experts exist in MoE-based LLMs?	In the era of Large Language Models (LLMs), the Mixture of Experts (MoE) architecture has emerged as an effective approach for training extremely large models with improved computational efficiency. This success builds upon extensive prior research aimed at enhancing expert specialization in MoE-based LLMs. However, th...	[ "Giang Do", "Hung Le", "Truyen Tran" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-04-07T00:00:00	https://arxiv.org/abs/2604.05267	https://arxiv.org/pdf/2604.05267v1	2604.05267	10.48550/arXiv.2604.05267	0	0	true	https://github.com/giangdip2410/Domain-specific-Experts	arXiv.org	0.827
99c67625df6867a7c38f19fa64c500d57e625dbe523bcb56c049dec8248e4854	[ "arxiv", "semantic_scholar" ]	QA-MoE: Towards a Continuous Reliability Spectrum with Quality-Aware Mixture of Experts for Robust Multimodal Sentiment Analysis	Multimodal Sentiment Analysis (MSA) aims to infer human sentiment from textual, acoustic, and visual signals. In real-world scenarios, however, multimodal inputs are often compromised by dynamic noise or modality missingness. Existing methods typically treat these imperfections as discrete cases or assume fixed corrupt...	[ "Yitong Zhu", "Yuxuan Jiang", "Guanxuan Jiang", "Bojing Hou", "Peng Yuan Zhou", "Ge Lin Kan", "Yuyang Wang" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-04-07T00:00:00	https://arxiv.org/abs/2604.05704	https://arxiv.org/pdf/2604.05704v2	2604.05704	10.48550/arXiv.2604.05704	0	0	false	null	arXiv.org	0.5351
08d93cdbbaa584f81f3ef8db0bc1f1dad5b9a94dacec6141ca75291d35b5f7a8	[ "arxiv", "semantic_scholar" ]	HI-MoE: Hierarchical Instance-Conditioned Mixture-of-Experts for Object Detection	Mixture-of-Experts (MoE) architectures enable conditional computation by activating only a subset of model parameters for each input. Although sparse routing has been highly effective in language models and has also shown promise in vision, most vision MoE methods operate at the image or patch level. This granularity i...	[ "Vadim Vashkelis", "Natalia Trukhina" ]	[ "cs.LG" ]	[ "Computer Science" ]	2026-04-06T00:00:00	https://arxiv.org/abs/2604.04908	https://arxiv.org/pdf/2604.04908v1	2604.04908	10.48550/arXiv.2604.04908	2	0	false	null	arXiv.org	0.534
4ed89d2d1ab879e6a0e4f4d9e4b00f8c2c37e8027b8c4ba572297b012bbc5f24	[ "arxiv", "semantic_scholar" ]	Three Phases of Expert Routing: How Load Balance Evolves During Mixture-of-Experts Training	We model Mixture-of-Experts (MoE) token routing as a congestion game with a single effective parameter, the congestion coefficient gamma_eff, that quantifies the balance-quality tradeoff. Tracking gamma_eff across training checkpoints of two open-source MoE models, OLMoE-1B-7B (20 checkpoints, with dense sampling in th...	[ "Charafeddine Mouzouni" ]	[ "cs.LG", "cs.AI", "cs.MA" ]	[ "Computer Science" ]	2026-04-05T00:00:00	https://arxiv.org/abs/2604.04230	https://arxiv.org/pdf/2604.04230v1	2604.04230	10.48550/arXiv.2604.04230	2	0	true	null	arXiv.org	0.8234
e39376cc2cf6add9fd4a108cdd97193c20af219bca92df878c46ae5b056ee2bf	[ "arxiv", "semantic_scholar" ]	Unveiling Language Routing Isolation in Multilingual MoE Models for Interpretable Subnetwork Adaptation	Mixture-of-Experts (MoE) models exhibit striking performance disparities across languages, yet the internal mechanisms driving these gaps remain poorly understood. In this work, we conduct a systematic analysis of expert routing patterns in MoE models, revealing a phenomenon we term Language Routing Isolation, in which...	[ "Kening Zheng", "Wei-Chieh Huang", "Jiahao Huo", "Zhonghao Li", "Henry Peng Zou", "Yibo Yan", "Xin Zou", "Jungang Li", "Junzhuo Li", "Hanrong Zhang", "Xuming Hu", "Philip S. Yu" ]	[ "cs.CL", "cs.AI" ]	[ "Computer Science" ]	2026-04-04T00:00:00	https://arxiv.org/abs/2604.03592	https://arxiv.org/pdf/2604.03592v1	2604.03592	10.48550/arXiv.2604.03592	2	2	false	null	arXiv.org	0.5317
a62dab8eb8e64f4c5e996644409c8b60bd38b04c9876d49cf6b00644f12629f7	[ "arxiv", "semantic_scholar" ]	The Expert Strikes Back: Interpreting Mixture-of-Experts Language Models at Expert Level	Mixture-of-Experts (MoE) architectures have become the dominant choice for scaling Large Language Models (LLMs), activating only a subset of parameters per token. While MoE architectures are primarily adopted for computational efficiency, it remains an open question whether their sparsity makes them inherently easier t...	[ "Jeremy Herbst", "Stefan Wermter", "Jae Hee Lee" ]	[ "cs.CL", "cs.AI", "cs.LG" ]	[ "Computer Science" ]	2026-04-02T00:00:00	https://arxiv.org/abs/2604.02178	https://arxiv.org/pdf/2604.02178v2	2604.02178	10.48550/arXiv.2604.02178	5	0	true	https://github.com/jerryy33/MoE_analysis	arXiv.org	0.8181
1d2e9986934d368ad386fc2e878f262e2c5616acbfca731dceb5fe1e507ad859	[ "arxiv", "semantic_scholar" ]	Routing-Free Mixture-of-Experts	Standard Mixture-of-Experts (MoE) models rely on centralized routing mechanisms that introduce rigid inductive biases. We propose Routing-Free MoE which eliminates any hard-coded centralized designs including external routers, Softmax, Top-K and load balancing, instead encapsulating all activation functionalities withi...	[ "Yilun Liu", "Jinru Han", "Sikuan Yan", "Volker Tresp", "Yunpu Ma" ]	[ "cs.LG", "cs.AI", "cs.CL" ]	[ "Computer Science" ]	2026-04-01T00:00:00	https://arxiv.org/abs/2604.00801	https://arxiv.org/pdf/2604.00801v1	2604.00801	10.48550/arXiv.2604.00801	0	0	true	https://github.com/liuyilun2000/RoutingFreeMoE/tree/release	arXiv.org	0.8164
edaf6c3f5f0ae614f6f446e377f12a44acff8afe7b6bf0e568b41ba819bd19cc	[ "arxiv", "semantic_scholar" ]	Self-Routing: Parameter-Free Expert Routing from Hidden States	Mixture-of-Experts (MoE) layers increase model capacity by activating only a small subset of experts per token, and typically rely on a learned router to map hidden states to expert assignments. In this work, we ask whether a dedicated learned router is strictly necessary in the MoE settings we study. We propose Self-R...	[ "Jama Hussein Mohamud", "Drew Wagner", "Mirco Ravanelli" ]	[ "cs.AI" ]	[ "Computer Science" ]	2026-04-01T00:00:00	https://arxiv.org/abs/2604.00421	https://arxiv.org/pdf/2604.00421v1	2604.00421	10.48550/arXiv.2604.00421	1	0	false	null	arXiv.org	0.5282
89337f2757e2c199bdaa57890555145ca8fbdba1f0e42d3e9604905d40e4450f	[ "arxiv", "semantic_scholar" ]	Routing Sensitivity Without Controllability: A Diagnostic Study of Fairness in MoE Language Models	Mixture-of-Experts (MoE) language models are universally sensitive to demographic content at the routing level, yet exploiting this sensitivity for fairness control is structurally limited. We introduce Fairness-Aware Routing Equilibrium (FARE), a diagnostic framework designed to probe the limits of routing-level stere...	[ "Junhyeok Lee", "Kyu Sung Choi" ]	[ "cs.CL" ]	[ "Computer Science" ]	2026-03-28T00:00:00	https://arxiv.org/abs/2603.27141	https://arxiv.org/pdf/2603.27141v1	2603.27141	10.48550/arXiv.2603.27141	0	0	false	null	arXiv.org	0.5236
8bcb27a67eed753bc3eb3f9de560c4155357e455dea1771114e21dbd073179b4	[ "arxiv", "semantic_scholar" ]	MoE-GRPO: Optimizing Mixture-of-Experts via Reinforcement Learning in Vision-Language Models	Mixture-of-Experts (MoE) has emerged as an effective approach to reduce the computational overhead of Transformer architectures by sparsely activating a subset of parameters for each token while preserving high model capacity. This paradigm has recently been extended to Vision-Language Models (VLMs), enabling scalable ...	[ "Dohwan Ko", "Jinyoung Park", "Seoung Choi", "Sanghyeok Lee", "Seohyun Lee", "Hyunwoo J. Kim" ]	[ "cs.CV" ]	[ "Computer Science" ]	2026-03-26T00:00:00	https://arxiv.org/abs/2603.24984	https://arxiv.org/pdf/2603.24984v2	2603.24984	10.48550/arXiv.2603.24984	1	0	false	null	arXiv.org	0.5214

End of preview. Expand in Data Studio

Mixture of Experts Papers — FineSet

A research-paper dataset on Mixture of Experts Papers, assembled, deduplicated, and quality-scored by FineSet from arXiv and Semantic Scholar.

📸 This is a dated snapshot — generated 2026-06-19. It is not auto-updated. Research on Mixture of Experts Papers moves fast — new papers land on arXiv every week. Want this same dataset refreshed daily, on a topic you choose? See the bottom. ↓

Why this dataset

Quality-scored: quality_score float (0–1), blends citations with recency + code/venue signals — filter out the noise
Papers with code: 120 flagged via has_code — find reproducible work fast
Deduplicated: arXiv + Semantic Scholar cross-referenced, duplicate records merged
Clean JSONL: 490 records, one per line, normalized fields — no encoding garbage

Dataset details

Records: 490
Date range: 2021–2026
Snapshot date: 2026-06-19 (frozen — see note above)
Sources: arXiv, Semantic Scholar (cross-referenced, duplicates merged)
arXiv categories: cs.LG, cs.CL
Quality scoring: citations + recency + code/venue blend, 0–1 (p50=0.35, p90=0.65)
Format: JSONL, one record per line

Fields

Field	Type	Description
id	string	Deterministic SHA256 record id
sources	list	Which sources contributed (`arxiv`, `semantic_scholar`)
title	string	Paper title
abstract	string	Full abstract
authors	list	Author names
categories	list	arXiv category codes
fields_of_study	list	Semantic Scholar field tags
published_date	string	ISO 8601 date
url	string	arXiv abstract URL
pdf_url	string\|null	Open-access PDF if available
arxiv_id	string\|null	arXiv identifier
doi	string\|null	DOI if available
citation_count	int	Citation count (Semantic Scholar)
influential_citation_count	int	Influential citations (Semantic Scholar)
has_code	bool	Code repo detected in the arXiv comment
code_url	string\|null	GitHub URL if detected
venue	string\|null	Publication venue
quality_score	float	0–1, blended (citations + recency + code/venue)

Quality score methodology

quality_score = max(impact, freshness), clamped to [0, 1], where:

impact = max( log10(citations+1)/4 , log10(influential_citations+1)/2 ) — realized impact (0.5 at 100 citations, ~0.75 at 1,000, 1.0 at 10,000+).
freshness = recency × (0.35 + 0.30·has_code + 0.20·has_venue) — a baseline for recent papers (so a strong paper published this week isn't scored 0 just for lacking citations), where recency is 1.0 for papers ≤60 days old and decays linearly to 0 by ~18 months.

Old highly-cited papers score on impact; brand-new papers score on freshness; old uncited papers score ~0. Useful for filtering training data by quality, not just age.

👉 Want this on YOUR topic, updated daily?

This snapshot is frozen at 2026-06-19. The live FineSet pipeline keeps a dataset like this refreshed every day on whatever topic you describe — new papers in, dedup and quality scoring automatic, export as JSONL/Parquet or push straight to the Hub.

Tell me the topic you'd want and I'll run the pipeline on it — open a discussion on this dataset, it's free and it's how I decide what to build next.

→ fineset.io — describe what you want to train on, get a dataset. Early-access waitlist open (referral skip available).

Downloads last month: 30