{"total":73,"rows":[{"id":"2501.01005","title":"FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving","created":"2025-01-02","cited_by_count":258,"area":"inference"},{"id":"2410.02367","title":"SageAttention: Accurate 8-Bit Attention for Plug-and-play Inference Acceleration","created":"2024-10-03","cited_by_count":160,"area":"inference"},{"id":"2407.18003","title":"Keep the Cost Down: A Review on Methods to Optimize LLM' s KV-Cache Consumption","created":"2024-07-25","cited_by_count":97,"area":"inference"},{"id":"2501.16383","title":"RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations","created":"2025-01-25","cited_by_count":46,"area":"inference"},{"id":"2409.17422","title":"Discovering the Gems in Early Layers: Accelerating Long-Context LLMs with 1000x Input Token Reduction","created":"2024-09-25","cited_by_count":39,"area":"inference"},{"id":"2502.10424","title":"QuantSpec: Self-Speculative Decoding with Hierarchical Quantized KV Cache","created":"2025-02-05","cited_by_count":27,"area":"inference"},{"id":"2503.18292","title":"Jenga: Effective Memory Management for Serving LLM with Heterogeneity","created":"2025-03-24","cited_by_count":25,"area":"inference"},{"id":"2602.04804","title":"OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models","created":"2026-02-04","cited_by_count":25,"area":"inference"},{"id":"2504.20068","title":"JITServe: SLO-aware LLM Serving with Imprecise Request Information","created":"2025-04-24","cited_by_count":17,"area":"inference"},{"id":"2603.18567","title":"SpecForge: A Flexible and Efficient Open-Source Training Framework for Speculative Decoding","created":"2026-03-19","cited_by_count":16,"area":"inference"},{"id":"2410.01805","title":"Locret: Enhancing Eviction in Long-Context LLM Inference with Trained Retaining Heads on Consumer-Grade Devices","created":"2024-10-02","cited_by_count":15,"area":"inference"},{"id":"2601.11580","title":"Speculative Decoding: Performance or Illusion?","created":"2025-12-31","cited_by_count":15,"area":"inference"},{"id":"2603.03251","title":"Speculative Speculative Decoding","created":"2026-03-03","cited_by_count":15,"area":"inference"},{"id":"2502.01068","title":"FastKV: Decoupling of Context Reduction and KV Cache Compression for Prefill-Decoding Acceleration","created":"2025-02-03","cited_by_count":14,"area":"inference"},{"id":"2509.18362","title":"FastMTP: Accelerating LLM Inference with Enhanced Multi-Token Prediction","created":"2025-09-16","cited_by_count":13,"area":"inference"},{"id":"2606.08382","title":"STAR-KV: Low-Rank KV Cache Compression via Soft Thresholding for Adaptive Rank Control","created":"2026-06-07","cited_by_count":11,"area":"inference"},{"id":"2505.21487","title":"Hardware-Efficient Attention for Fast Decoding","created":"2025-05-27","cited_by_count":10,"area":"inference"},{"id":"2502.12574","title":"HeadInfer: Memory-Efficient LLM Inference by Head-wise Offloading","created":"2025-02-18","cited_by_count":9,"area":"inference"},{"id":"2507.14204","title":"LaCache: Ladder-Shaped KV Caching for Efficient Long-Context Modeling of Large Language Models","created":"2025-07-14","cited_by_count":9,"area":"inference"},{"id":"2605.17757","title":"OSCAR: Offline Spectral Covariance-Aware Rotation for 2-bit KV Cache Quantization","created":"2026-05-18","cited_by_count":9,"area":"inference"},{"id":"2503.18773","title":"BitDecoding: Unlocking Tensor Cores for Long-Context LLMs with Low-Bit KV Cache","created":"2025-03-24","cited_by_count":8,"area":"inference"},{"id":"2509.01229","title":"LiquidGEMM: Hardware-Efficient W4A8 GEMM Kernel for High-Performance LLM Serving","created":"2025-09-01","cited_by_count":8,"area":"inference"},{"id":"2603.17435","title":"ZipServ: Fast and Memory-Efficient LLM Inference with Hardware-Aware Lossless Compression","created":"2026-03-18","cited_by_count":8,"area":"inference"},{"id":"2602.07223","title":"Vegas: Self-Speculative Decoding with Verification-Guided Sparse Attention","created":"2026-02-06","cited_by_count":7,"area":"inference"},{"id":"2606.06256","title":"RedKnot: Efficient Long-Context LLM Serving with Head-Aware KV Reuse and SegPagedAttention","created":"2026-06-04","cited_by_count":7,"area":"inference"},{"id":"2502.15572","title":"DReSD: Dense Retrieval for Speculative Decoding","created":"2025-02-21","cited_by_count":6,"area":"inference"},{"id":"2504.14992","title":"Efficient Pretraining Length Scaling","created":"2025-04-21","cited_by_count":6,"area":"inference"},{"id":"2502.15955","title":"Compression Barriers for Autoregressive Transformers","created":"2025-02-21","cited_by_count":5,"area":"inference"},{"id":"2504.14775","title":"gLLM: Global Balanced Pipeline Parallelism System for Distributed LLM Serving with Token Throttling","created":"2025-04-21","cited_by_count":5,"area":"inference"},{"id":"2511.22333","title":"PAT: Accelerating LLM Decoding via Prefix-Aware Attention with Resource Efficient Multi-Tile Kernel","created":"2025-11-27","cited_by_count":5,"area":"inference"},{"id":"2502.15779","title":"Rotate, Clip, and Partition: Towards W2A4KV4 Quantization by Integrating Rotation and Learnable Non-uniform Quantizer","created":"2025-02-17","cited_by_count":4,"area":"inference"},{"id":"2504.06319","title":"Accelerating LLM Inference Throughput via Asynchronous KV Cache Prefetching","created":"2025-04-08","cited_by_count":4,"area":"inference"},{"id":"2510.06175","title":"VecInfer: Efficient LLM Inference with Low-Bit KV Cache via Outlier-Suppressed Vector Quantization","created":"2025-10-07","cited_by_count":4,"area":"inference"},{"id":"2511.04791","title":"DuetServe: Harmonizing Prefill and Decode for LLM Serving via Adaptive GPU Multiplexing","created":"2025-11-06","cited_by_count":4,"area":"inference"},{"id":"2511.16193","title":"Fast LLM Post-training via Decoupled and Fastest-of-N Speculation","created":"2025-11-20","cited_by_count":4,"area":"inference"},{"id":"2605.30571","title":"Memory-Bound but Not Bandwidth-Limited: The Physical AI Inference Gap in Batch-1 LLM Decode","created":"2026-05-28","cited_by_count":4,"area":"inference"},{"id":"2510.13223","title":"BanaServe: Unified KV Cache and Dynamic Module Migration for Balancing Disaggregated LLM Serving in AI Infrastructure","created":"2025-10-15","cited_by_count":3,"area":"inference"},{"id":"2606.23581","title":"Kamera: Unified Position-Invariant Multimodal KV Cache for Training-Free Reuse","created":"2026-06-22","cited_by_count":3,"area":"inference"},{"id":"2505.20839","title":"FireQ: Fast INT4-FP8 Kernel and RoPE-aware Quantization for LLM Inference Acceleration","created":"2025-05-27","cited_by_count":2,"area":"inference"},{"id":"2509.04827","title":"VoltanaLLM: Energy-Efficient and SLO-Aware Disaggregated LLM Serving via Adaptive Frequency Control and State-Space Routing","created":"2025-09-05","cited_by_count":2,"area":"inference"},{"id":"2603.02188","title":"Multi-Head Low-Rank Attention","created":"2026-03-02","cited_by_count":2,"area":"inference"},{"id":"2603.04427","title":"Thin Keys, Full Values: Reducing KV Cache via Low-Dimensional Attention Selection","created":"2026-02-16","cited_by_count":2,"area":"inference"},{"id":"2604.05718","title":"MPM: Mutual Pair Merging for Efficient Vision Transformers","created":"2026-04-07","cited_by_count":2,"area":"inference"},{"id":"2605.21226","title":"OCTOPUS: Optimized KV Cache for Transformers via Octahedral Parametrization Under optimal Squared error quantization","created":"2026-05-20","cited_by_count":2,"area":"inference"},{"id":"2606.21023","title":"Demystifying Numerical Instability in LLM Inference: Achieving Reproducible Inference for Mission-Critical Tasks with HEAL","created":"2026-06-19","cited_by_count":2,"area":"inference"},{"id":"2608.19758","title":"FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving","created":"2026-08-20","cited_by_count":2,"area":"inference"},{"id":"2412.18106","title":"Tackling the Dynamicity in a Production LLM Serving System with SOTA Optimizations via Hybrid Prefill/Decode/Verify Scheduling on Efficient Meta-kernels","created":"2024-12-24","cited_by_count":1,"area":"inference"},{"id":"2510.27641","title":"SpecAttn: Speculating Sparse Attention","created":"2025-10-31","cited_by_count":1,"area":"inference"},{"id":"2603.12038","title":"Slow-Fast Inference: Training-Free Inference Acceleration via Within-Sentence Support Stability","created":"2026-03-12","cited_by_count":1,"area":"inference"},{"id":"2604.15464","title":"Ragged Paged Attention: A High-Performance and Flexible LLM Inference Kernel for TPU","created":"2026-04-16","cited_by_count":1,"area":"inference"},{"id":"2605.11999","title":"The Illusion of Power Capping in LLM Decode: A Phase-Aware Energy Characterisation Across Attention Architectures","created":"2026-05-12","cited_by_count":1,"area":"inference"},{"id":"2607.00760","title":"MosaicKV: Serving Long-Context LLM with Dynamic Two-D KV Cache Compression","created":"2026-07-01","cited_by_count":1,"area":"inference"},{"id":"2608.02947","title":"ATFlash: Per-RoPE-Wavelength Attention Windows for Compute/Memory-Efficient LLM Inference","created":"2026-08-03","cited_by_count":1,"area":"inference"},{"id":"2609.02737","title":"Language Models Can Control Their Own Attention","created":"2026-09-02","cited_by_count":1,"area":"inference"},{"id":"2609.36173","title":"Draft in Parallel, Condition Through Depth: Adjacent Causal Injection for Speculative Decoding","created":"2026-09-28","cited_by_count":1,"area":"inference"},{"id":"2509.23384","title":"A Predictive and Synergistic Two-Layer Scheduling Framework for LLM Serving","created":"2025-09-27","cited_by_count":0,"area":"inference"},{"id":"2511.12031","title":"Striking the Right Balance between Compute and Copy: Improving LLM Inferencing Under Speculative Decoding","created":"2025-11-15","cited_by_count":0,"area":"inference"},{"id":"2603.04797","title":"Hardware-Software Co-design for 3D-DRAM-based LLM Serving Accelerator","created":"2026-03-05","cited_by_count":0,"area":"inference"},{"id":"2604.05942","title":"BOSCH: Black-Box Binary Optimization for Short-Context Attention-Head Selection in LLMs","created":"2026-04-07","cited_by_count":0,"area":"inference"},{"id":"2605.11317","title":"SOMA: Efficient Multi-turn LLM Serving via Small Language Model","created":"2026-05-11","cited_by_count":0,"area":"inference"},{"id":"2606.26666","title":"PersistentKV: Page-Aware Decode Scheduling for Long-Context LLM Serving on Commodity GPUs","created":"2026-06-25","cited_by_count":0,"area":"inference"},{"id":"2607.19349","title":"FineServe: A Fine-Grained Dataset and Characterization of Global LLM Serving Workloads","created":"2026-04-17","cited_by_count":0,"area":"inference"},{"id":"2608.17336","title":"TileMix: Tile-Centric Mixed-Precision Attention for LLM Inference Acceleration","created":"2026-08-18","cited_by_count":0,"area":"inference"},{"id":"2609.13205","title":"Self-Indexing Attention for Compression-Compatible Sparse Long-Context LLM Inference","created":"2026-08-16","cited_by_count":0,"area":"inference"},{"id":"2609.16491","title":"PipeSwift: Revisiting Pipeline Parallelism for Large-Scale Completion-Oriented Agentic LLM Serving","created":"2026-09-15","cited_by_count":0,"area":"inference"},{"id":"2609.17008","title":"FlexEE: Self-Speculative and KV-Compatible Early Exiting for Offloading-Aware LLM Inference","created":"2026-09-15","cited_by_count":0,"area":"inference"},{"id":"2609.21172","title":"TierKV: Long-Context On-Device LLMs via Predictive Multi-Tier KV Caching","created":"2026-09-18","cited_by_count":0,"area":"inference"},{"id":"2609.23130","title":"From Inference Engine to Inference Control Plane: Connecting vLLM, llm-d, and the Evolution of Efficient Distributed LLM Serving","created":"2026-09-19","cited_by_count":0,"area":"inference"},{"id":"2609.26796","title":"Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs","created":"2026-09-22","cited_by_count":0,"area":"inference"},{"id":"2609.33410","title":"FoldAttention: Declared-Reference Softmax for Fast Decode and Deterministic Backward","created":"2026-09-27","cited_by_count":0,"area":"inference"},{"id":"2609.34653","title":"OmniTide: Co-Designing Algorithms and Systems for Efficient On-Device Omni-LLM Streaming","created":"2026-09-28","cited_by_count":0,"area":"inference"},{"id":"2610.04313","title":"MOIRA: Mass-Oriented Indexing with Ragged Attention for Long-Context Decoding","created":"2026-10-03","cited_by_count":0,"area":"inference"},{"id":"2610.09051","title":"A Self-Pruning Transformer: Extreme KV-Cache Compression with Universal Attention","created":"2026-10-06","cited_by_count":0,"area":"inference"}],"next_cursor":null}