{"total":80,"rows":[{"id":"2502.13189","title":"MoBA: Mixture of Block Attention for Long-Context LLMs","created":"2025-02-18","cited_by_count":198,"area":"moe"},{"id":"2410.13276","title":"SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs","created":"2024-10-17","cited_by_count":124,"area":"moe"},{"id":"2411.11217","title":"MoE-Lightning: High-Throughput MoE Inference on Memory-constrained GPUs","created":"2024-11-18","cited_by_count":73,"area":"moe"},{"id":"2308.14352","title":"EdgeMoE: Empowering Sparse Large Language Models on Mobile Devices","created":"2023-08-28","cited_by_count":69,"area":"moe"},{"id":"2405.16039","title":"MoEUT: Mixture-of-Experts Universal Transformers","created":"2024-05-25","cited_by_count":59,"area":"moe"},{"id":"2410.22134","title":"ProMoE: Fast MoE-based LLM Serving using Proactive Caching","created":"2024-10-29","cited_by_count":40,"area":"moe"},{"id":"2505.09343","title":"Insights into DeepSeek-V3: Scaling Challenges and Reflections on Hardware for AI Architectures","created":"2025-05-14","cited_by_count":39,"area":"moe"},{"id":"2505.22323","title":"Advancing Expert Specialization for Better MoE","created":"2025-05-28","cited_by_count":39,"area":"moe"},{"id":"2312.07987","title":"SwitchHead: Accelerating Transformers with Mixture-of-Experts Attention","created":"2023-12-13","cited_by_count":34,"area":"moe"},{"id":"2404.05019","title":"Shortcut-connected Expert Parallelism for Accelerating Mixture-of-Experts","created":"2024-04-07","cited_by_count":32,"area":"moe"},{"id":"2504.02263","title":"MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism","created":"2025-04-03","cited_by_count":32,"area":"moe"},{"id":"2505.11432","title":"MegaScale-MoE: Large-Scale Communication-Efficient Training of Mixture-of-Experts Models in Production","created":"2025-05-16","cited_by_count":31,"area":"moe"},{"id":"2508.02317","title":"VeOmni: Scaling Any Modality Model Training with Model-Centric Distributed Recipe Zoo","created":"2025-08-04","cited_by_count":20,"area":"moe"},{"id":"2408.11396","title":"MoE-LPR: Multilingual Extension of Large Language Models through Mixture-of-Experts with Language Priors Routing","created":"2024-08-21","cited_by_count":16,"area":"moe"},{"id":"2505.00315","title":"Mixture of Sparse Attention: Content-Based Learnable Sparse Attention via Expert-Choice Routing","created":"2025-05-01","cited_by_count":14,"area":"moe"},{"id":"2506.04667","title":"FlashMoE: Fast Distributed MoE in a Single Kernel","created":"2025-06-05","cited_by_count":14,"area":"moe"},{"id":"2502.03009","title":"Scaling Laws for Upcycling Mixture-of-Experts Language Models","created":"2025-02-05","cited_by_count":13,"area":"moe"},{"id":"2508.15036","title":"MoEcho: Exploiting Side-Channel Attacks to Compromise User Privacy in Mixture-of-Experts LLMs","created":"2025-08-20","cited_by_count":13,"area":"moe"},{"id":"2506.07533","title":"MoQAE: Mixed-Precision Quantization for Long-Context LLM Inference via Mixture of Quantization-Aware Experts","created":"2025-06-09","cited_by_count":12,"area":"moe"},{"id":"2407.11030","title":"DLO: Dynamic Layer Operation for Efficient Vertical Scaling of LLMs","created":"2024-07-03","cited_by_count":10,"area":"moe"},{"id":"2602.03560","title":"HySparse: A Hybrid Sparse Attention Architecture with Oracle Token Selection and KV Cache Sharing","created":"2026-02-03","cited_by_count":10,"area":"moe"},{"id":"2506.05426","title":"Mixture-of-Experts Meets In-Context Reinforcement Learning","created":"2025-06-05","cited_by_count":8,"area":"moe"},{"id":"2508.21706","title":"Accelerating Mixture-of-Experts Inference by Hiding Offloading Latency with Speculative Decoding","created":"2025-08-29","cited_by_count":8,"area":"moe"},{"id":"2504.03871","title":"HeterMoE: Efficient Training of Mixture-of-Experts Models on Heterogeneous GPUs","created":"2025-04-04","cited_by_count":7,"area":"moe"},{"id":"2412.02503","title":"VA-MoE: Variables-Adaptive Mixture of Experts for Incremental Weather Forecasting","created":"2024-12-03","cited_by_count":6,"area":"moe"},{"id":"2508.06526","title":"PiKV: KV Cache Management System for Mixture of Experts","created":"2025-08-02","cited_by_count":5,"area":"moe"},{"id":"2604.19241","title":"UniEP: Unified Expert-Parallel MoE MegaKernel for LLM Training","created":"2026-04-21","cited_by_count":5,"area":"moe"},{"id":"2408.04307","title":"MoC-System: Efficient Fault Tolerance for Sparse Mixture-of-Experts Model Training","created":"2024-08-08","cited_by_count":4,"area":"moe"},{"id":"2502.00281","title":"Sigmoid Self-Attention has Lower Sample Complexity than Softmax Self-Attention: A Mixture-of-Experts Perspective","created":"2025-02-01","cited_by_count":4,"area":"moe"},{"id":"2506.07366","title":"MoE-GPS: Guidlines for Prediction Strategy for Dynamic Expert Duplication in MoE Load Balancing","created":"2025-06-09","cited_by_count":4,"area":"moe"},{"id":"2507.15465","title":"Rethinking LLM Inference Bottlenecks: Insights from Latent Attention and Mixture-of-Experts","created":"2025-07-21","cited_by_count":4,"area":"moe"},{"id":"2507.17133","title":"BrownoutServe: SLO-Aware Inference Serving under Bursty Workloads for MoE-based LLMs","created":"2025-07-23","cited_by_count":4,"area":"moe"},{"id":"2510.08055","title":"From Tokens to Layers: Redefining Stall-Free Scheduling for MoE Serving with Layered Prefill","created":"2025-10-09","cited_by_count":4,"area":"moe"},{"id":"2504.09345","title":"MoE-Lens: Towards the Hardware Limit of High-Throughput MoE LLM Serving Under Resource Constraints","created":"2025-04-12","cited_by_count":3,"area":"moe"},{"id":"2506.16419","title":"Optimizing MoE Routers: Design, Implementation, and Evaluation in Transformer Models","created":"2025-06-19","cited_by_count":3,"area":"moe"},{"id":"2509.24406","title":"Muon: Training and Trade-offs with Latent Attention and MoE","created":"2025-09-29","cited_by_count":3,"area":"moe"},{"id":"2509.25689","title":"Collaborative Compression for Large-Scale MoE Deployment on Edge","created":"2025-09-30","cited_by_count":3,"area":"moe"},{"id":"2601.21641","title":"Seg-MoE: Multi-Resolution Segment-wise Mixture-of-Experts for Time Series Forecasting Transformers","created":"2026-01-29","cited_by_count":3,"area":"moe"},{"id":"2602.22059","title":"NESTOR: A Nested MOE-based Neural Operator for Large-Scale PDE Pre-Training","created":"2026-02-25","cited_by_count":3,"area":"moe"},{"id":"2607.12696","title":"Less Experts, Faster Decoding: Cost-Aware Speculative Decoding for Mixture-of-Experts","created":"2026-07-14","cited_by_count":3,"area":"moe"},{"id":"2503.10725","title":"Samoyeds: Accelerating MoE Models with Structured Sparsity Leveraging Sparse Tensor Cores","created":"2025-03-13","cited_by_count":2,"area":"moe"},{"id":"2508.16700","title":"GPT-OSS-20B: A Comprehensive Deployment-Centric Analysis of OpenAI's Open-Weight Mixture of Experts Model","created":"2025-08-22","cited_by_count":2,"area":"moe"},{"id":"2510.13079","title":"GatePro: Parameter-Free Expert Selection Optimization for Mixture-of-Experts Models","created":"2025-10-15","cited_by_count":2,"area":"moe"},{"id":"2510.18825","title":"Unifying and Enhancing Graph Transformers via a Hierarchical Mask Framework","created":"2025-10-21","cited_by_count":2,"area":"moe"},{"id":"2601.11564","title":"Context Discipline and Performance Correlation: Analyzing LLM Performance and Quality Degradation Under Varying Context Lengths","created":"2025-12-25","cited_by_count":2,"area":"moe"},{"id":"2602.01219","title":"Mixture-of-Top-k Attention: Efficient Attention via Scalable Fast Weights","created":"2026-02-01","cited_by_count":2,"area":"moe"},{"id":"2602.04870","title":"Multi-Head LatentMoE and Head Parallel: Communication-Efficient and Deterministic MoE Parallelism","created":"2026-02-04","cited_by_count":2,"area":"moe"},{"id":"2605.02960","title":"MoE-Prefill: Zero Redundancy Overheads in MoE Prefill Serving","created":"2026-05-03","cited_by_count":2,"area":"moe"},{"id":"2503.02495","title":"Union of Experts: Adapting Hierarchical Routing to Equivalently Decomposed Transformer","created":"2025-03-04","cited_by_count":1,"area":"moe"},{"id":"2510.08008","title":"Beyond Sunk Costs: Boosting LLM Pre-training Efficiency via Orthogonal Growth of Mixture-of-Experts","created":"2025-10-09","cited_by_count":1,"area":"moe"},{"id":"2603.07006","title":"Mozart: Modularized and Efficient MoE Training on 3.5D Wafer-Scale Chiplet Architectures","created":"2026-03-07","cited_by_count":1,"area":"moe"},{"id":"2603.19289","title":"Speculating Experts Accelerates Inference for Mixture-of-Experts","created":"2026-03-09","cited_by_count":1,"area":"moe"},{"id":"2606.24506","title":"CrossPool: Efficient Multi-LLM Serving for Cold MoE Models through KV-Cache and Weight Disaggregation","created":"2026-06-23","cited_by_count":1,"area":"moe"},{"id":"2606.26607","title":"Moebius: Serving Mixture-of-Expert Models with Seamless Runtime Parallelism Switch","created":"2026-06-25","cited_by_count":1,"area":"moe"},{"id":"2607.01844","title":"Mixture-of-Parallelisms: Towards Memory-Efficient Training Stack for Mixture-of-Experts Models","created":"2026-07-02","cited_by_count":1,"area":"moe"},{"id":"2607.20145","title":"SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD","created":"2026-07-22","cited_by_count":1,"area":"moe"},{"id":"2505.06481","title":"QoS-Efficient Serving of Multiple Mixture-of-Expert LLMs Using Partial Runtime Reconfiguration","created":"2025-05-10","cited_by_count":0,"area":"moe"},{"id":"2512.23029","title":"Viability and Performance of a Private LLM Server for SMBs: A Benchmark Analysis of Qwen3-30B on Consumer-Grade Hardware","created":"2025-12-28","cited_by_count":0,"area":"moe"},{"id":"2601.08215","title":"Towards Principled Design of Mixture-of-Experts Language Models under Memory and Inference Constraints","created":"2026-01-13","cited_by_count":0,"area":"moe"},{"id":"2601.21866","title":"MoHETS: Long-term Time Series Forecasting with Mixture-of-Heterogeneous-Experts","created":"2026-01-29","cited_by_count":0,"area":"moe"},{"id":"2602.11543","title":"Pretraining A Large Language Model using Distributed GPUs: A Memory-Efficient Decentralized Paradigm","created":"2026-02-12","cited_by_count":0,"area":"moe"},{"id":"2603.10031","title":"Architecture-Aware LLM Inference Optimization on AMD Instinct GPUs: A Comprehensive Benchmark and Deployment Study","created":"2026-02-27","cited_by_count":0,"area":"moe"},{"id":"2605.05049","title":"Piper: Efficient Large-Scale MoE Training via Resource Modeling and Pipelined Hybrid Parallelism","created":"2026-05-06","cited_by_count":0,"area":"moe"},{"id":"2605.18848","title":"Exact Linear Attention","created":"2026-05-13","cited_by_count":0,"area":"moe"},{"id":"2605.19945","title":"GEM: GPU-Variability-Aware Expert to GPU Mapping for MoE Systems","created":"2026-05-19","cited_by_count":0,"area":"moe"},{"id":"2605.23764","title":"HyperParallel-MoE: Multi-Core Interleaved Scheduling for Fast MoE Training on Ascend NPUs","created":"2026-05-22","cited_by_count":0,"area":"moe"},{"id":"2605.24681","title":"Mix-MoE: Improving Multilingual Machine Translation of Large Language Models through Mixed MoEs","created":"2026-05-23","cited_by_count":0,"area":"moe"},{"id":"2606.22541","title":"ASAP: A Disaggregated and Asynchronous Inference System for MoE Prefill","created":"2026-06-21","cited_by_count":0,"area":"moe"},{"id":"2607.06601","title":"TriRoute: Unified Learned Routing for Joint Adaptive Attention, Experts, and KV-Cache Allocation","created":"2026-07-07","cited_by_count":0,"area":"moe"},{"id":"2607.08215","title":"On the Limitations of Non-GPU AI Accelerators for Large-Model Inference: A Field Study of MoE and Multimodal Serving on Huawei Ascend","created":"2026-07-09","cited_by_count":0,"area":"moe"},{"id":"2608.06723","title":"Multi-Level Modeling of Large Language Model Inference Latency and Energy via Hybrid Analytical--Machine-Learning Predictors","created":"2026-08-07","cited_by_count":0,"area":"moe"},{"id":"2608.08081","title":"RotaryQuant: Fitting 120B MoE Models on Consumer Hardware via Fused Compressed-Space Attention","created":"2026-08-08","cited_by_count":0,"area":"moe"},{"id":"2608.10605","title":"Compute-Optimal Is Not Cluster-Optimal: Systems-Aware Scaling for Sparse Mixture-of-Experts","created":"2026-08-11","cited_by_count":0,"area":"moe"},{"id":"2609.04244","title":"MonoMoE: An Efficient Fused Mega-kernel for Quantized MoE Decoding","created":"2026-08-19","cited_by_count":0,"area":"moe"},{"id":"2609.08663","title":"MoEMB: Scaling Universal Multimodal Embeddings with Efficient Mixture-of-Experts Models","created":"2026-09-08","cited_by_count":0,"area":"moe"},{"id":"2609.09748","title":"Epoch: Compiling Diffusion Blocks for Sparse MoE Serving","created":"2026-09-09","cited_by_count":0,"area":"moe"},{"id":"2609.14306","title":"Flattening Every Memory Peak in Long-Context Mixture-of-Experts Training","created":"2026-09-13","cited_by_count":0,"area":"moe"},{"id":"2609.16179","title":"Z-Loss Backward Geometry in Dense Output Heads and Sparse Routers","created":"2026-07-31","cited_by_count":0,"area":"moe"},{"id":"2609.20974","title":"Attention-Aware Routing: Coupling Routing and Attention in MoEs","created":"2026-09-17","cited_by_count":0,"area":"moe"},{"id":"2610.16890","title":"FORGE: Procedural Weight Generation for Sparse MoE Inference","created":"2026-10-12","cited_by_count":0,"area":"moe"}],"next_cursor":null}