Key papers

• Less Is More: Optimizing Function Calling For LLM Execution On Edge Devices (2024)
  Varatheepan Paramanayakam, Andreas Karatzas, Iraklis Anagnostopoulos, et al.
  6.77
• MoE$^2$-LoRA: When MoE Models Meet MoE-style Low-Rank Adaptation (2026)
  Qingyu Yang et al.
  5.01
• Unified Static-Dynamic Pruning for Efficient LLM Inference (2026)
  Jinhyeok Kim et al.
  5.01
• The Wiola Architecture for Efficient Small Language Models (2026)
  Aryuemaan Kumar Chowdhury et al.
  4.39
• PEEK: Predictive Queue-Informed KV Cache Management for LLM Serving (2026)
  Bing Xie et al.
  4.39
• HCRMap: Pressure-Aware Hot-Expert Residency Mapping for 3.5D MoE Chiplet Inference (2026)
  Yongqin Zhang
  4.39
• Full-Pipeline Inference Optimization for MiMo-V2.5 Series: Pushing Hybrid SWA Efficiency to the Limit (2026)
  Xiaomi MiMo Team et al.
  4.39
• Beyond Single-Dimensional Compression: The Compound Sparsity Frontier of Large Language Models (2026)
  Chao Han et al.
  4.39
• Transferable Latency Prediction for Fast LLM Screening on Heterogeneous Edge Devices (2026)
  Xiaolong Tu et al.
  4.39
• Sparse by Command: Task-Conditional Compute Skipping for Multi-Task Inference Accelerators (2026)
  Afzal Ahmad et al.
  4.39
• \k{appa}-LoRA: Condition Numbers Reveal Which LoRA Matrices Worth Updating (2026)
  Jianghui Wang et al.
  4.39
• How Far Can Disaggregation Go? A Design-Space Exploration of Attention-FFN Disaggregation for Efficient MoE LLM Serving (2026)
  Hanjiang Wu et al.
  4.27
• Life Cycle Assessment of Pre-training the Lucie 7B Open-Source Large Language Model on the Jean Zay Supercomputer (2026)
  Marc L\'eobet et al.
  3.51
• PoLoRA: A Preconditioned Orthogonalized LoRA Optimizer (2026)
  Nikhil Ghosh et al.
  3.51
• Convolution for Large Language Models (2026)
  Yuchuan Tian et al.
  3.51
• SimpleTool: Parallel Decoding for Real-Time LLM Function Calling (2026)
  Xiao Shi et al.
  3.27
• Adaptive LoRA Merge with Parameter Pruning for Low-Resource Generation (2025)
  Ryota Miyano and Yuki Arase
  2.82
• Breaking the MoE LLM Trilemma: Dynamic Expert Clustering with Structured Compression (2025)
  Peijun Zhu et al.
  2.56
• LiMuon: Light and Fast Muon Optimizer for Large Models (2025)
  Feihu Huang et al.
  2.51
• Looped Latent Attention: Cross-Loop KV Compression for Looped Transformers (2026)
  James O' Neill et al.
  2.00
• WaveFilter: Enhancing the Long-Context Capability of Diffusion LLMs via Wavelet-Guided KV Cache Filtering (2026)
  Jinnan Yang et al.
  1.94
• Parameter-Efficient Fine-Tuning with Learnable Rank (2026)
  Arpit Garg et al.
  1.94
• Token-Operations-Oriented Inference Optimization Techniques for Large Models (2026)
  Shiguo Lian et al.
  1.94
• PACE: Prune-And-Compress Ensemble Models (2026)
  Fabian Akkerman et al.
  1.89
• TileQ: Efficient Low-Rank Quantization of Mixture-of-Experts with 2D Tiling (2026)
  Hongyaoxing Gu et al.
  1.89
• Freeze Deep, Train Shallow: Interpretable Layer Allocation for Continued Pre-Training (2026)
  Yu-Hang Wu et al.
  1.89
• TurboGR: An Accelerated Training System for Large-Scale Generative Recommendation (2026)
  Huichao Chai and Zhixin Wu and Xuemiao Li and Shiqing Fan and Hengfeng Wang and Maojun Peng and Lu Xu and Yaoyuan Wang and Yibo Jin and Wei Guo and Yongxiang Feng
  1.89
• BEAM: Binary Expert Activation Masking for Dynamic Routing in MoE (2026)
  Juntong Wu et al.
  1.89
• Ghosted Layers: Unconstrained Activation Alignment for Recovering Layer-Pruned LLMs (2026)
  Vincent-Daniel Yun et al.
  1.89
• Quant.npu: Enabling Efficient Mobile NPU Inference for on-device LLMs via Fully Static Quantization (2026)
  Jinghe Zhang et al.
  1.89
• Learning-Augmented Online Scheduling with Parsimonious Preemption (2026)
  Mugen Blue et al.
  1.89
• PromptEmbedder: Efficient and Transferable Text Embedding via Dual-LLM Soft Prompting (2026)
  Yu-Che Tsai et al.
  1.89
• In-Place Test-Time Training (2026)
  Guhao Feng et al.
  1.83
• New Hybrid Fine-Tuning Paradigm for LLMs: Algorithm Design and Convergence Analysis Framework (2026)
  Shaocong Ma et al.
  1.83
• A Layer-wise Analysis of Supervised Fine-Tuning (2026)
  Qinghua Zhao et al.
  1.83
• Polynomial Expansion Rank Adaptation: Enhancing Low-Rank Fine-Tuning with High-Order Interactions (2026)
  Wenhao Zhang and Lin Mu and Li Ni and Peiquan Jin and Yiwen Zhang
  1.83
• TLoRA+: A Low-Rank Parameter-Efficient Fine-Tuning Method for Large Language Models (2026)
  Yarui Cao et al.
  1.83
• VisPCO: Visual Token Pruning Configuration Optimization via Budget-Aware Pareto-Frontier Learning for Vision-Language Models (2026)
  Huawei Ji and Yuanhao Sun and Yuan Jin and Cheng Deng and Jiaxin Ding and Luoyi Fu and Xinbing Wang
  1.83
• SAMoRA: Semantic-Aware Mixture of LoRA Experts for Task-Adaptive Learning (2026)
  Boyan Shi et al.
  1.83
• CommFuse: Hiding Tail Latency via Communication Decomposition and Fusion for Distributed LLM Training (2026)
  Rezaul Karim et al.
  1.83
• AHASD: Asynchronous Heterogeneous Architecture for LLM Adaptive Drafting Speculative Decoding on Mobile Devices (2026)
  Ma Zirui et al.
  1.83
• Unbiased Dynamic Pruning for Efficient Group-Based Policy Optimization (2026)
  Haodong Zhu and Yangyang Ren and Yanjing Li and Mingbao Lin and Linlin Yang and Xuhui Liu and Xiantong Zhen and Haiguang Liu and Baochang Zhang
  1.78
• Expert Pyramid Tuning: Efficient Parameter Fine-Tuning for Expertise-Driven Task Allocation (2026)
  Jia-Chen Zhang et al.
  1.78
• NanoZK: Privacy-Preserving Verifiable Inference for Large Language Models via Layerwise Zero-Knowledge Proofs (2026)
  Zhaohui Wang
  1.78
• MarkovScale: Towards Optimal Sequential Scaling at Inference Time (2026)
  Youkang Wang et al.
  1.72
• NLI:Non-uniform Linear Interpolation Approximation of Nonlinear Operations for Efficient LLMs Inference (2026)
  Jiangyong Yu et al.
  1.72
• Layer-wise LoRA fine-tuning: a similarity metric approach (2026)
  Keith Ando Ogawa et al.
  1.72
• SHINE: A Scalable In-Context Hypernetwork for Mapping Context to LoRA in a Single Pass (2026)
  Yewei Liu et al.
  1.72
• Ruyi2 Technical Report (2026)
  Huan Song et al.
  1.72
• ADEPT: Adaptive Dynamic Early-Exit Process for Transformers (2026)
  Sangmin Yoo and Srikanth Malla and Chiho Choi and Wei D. Lu and Joon Hee Choi
  1.67
• Fast KVzip: Efficient and Accurate LLM Inference with Gated KV Eviction (2026)
  Jang-Hyun Kim et al.
  1.67
• GRASP: GRouped Activation Shared Parameterization for Parameter-Efficient Fine-Tuning and Robust Inference of Transformers (2025)
  Malyaban Bal and Abhronil Sengupta
  1.61
• Turbo-Muon: Accelerating Orthogonality-Based Optimization with Pre-Conditioning (2025)
  Thibaut Boissin (IRIT-MISFIT) et al.
  1.61
• Leveraging KV Similarity for Online Structured Pruning in LLMs (2025)
  Jungmin Lee et al.
  1.61
• CALM: A CKA-Guided Adaptive Layer-Wise Modularization Framework for LLM Quantization (2025)
  Jinhao Zhang and Yunquan Zhang and Daning Chen and JunSun and Zicheng Yan
  1.61
• MoE Pathfinder: Trajectory-driven Expert Pruning (2025)
  Xican Yang et al.
  1.61
• AFA-LoRA: Enabling Non-Linear Adaptations in LoRA with Activation Function Annealing (2025)
  Jiacheng Li et al.
  1.61
• Yggdrasil: Bridging Dynamic Speculation and Static Runtime for Latency-Optimal Tree-Based LLM Decoding (2025)
  Yue Guan et al.
  1.61
• Training Report of TeleChat3-MoE (2025)
  Xinzhang Liu et al.
  1.61
• Loquetier: A Virtualized Multi-LoRA Framework for Unified LLM Fine-tuning and Serving (2025)
  Yuchen Zhang et al.
  1.56