English
Related papers

Related papers: When Hidden States Drift: Can KV Caches Rescue Lon…

200 papers

The Key-Value (KV) cache is a crucial component in serving transformer-based autoregressive large language models (LLMs), enabling faster inference by storing previously computed KV vectors. However, its memory consumption scales linearly…

Machine Learning · Computer Science 2024-10-07 Rongzhi Zhang , Kuang Wang , Liyuan Liu , Shuohang Wang , Hao Cheng , Chao Zhang , Yelong Shen

Speculative decoding accelerates large language model inference by pairing a target model with a lightweight draft model whose proposed tokens are verified in parallel. A common way to build draft models, like EAGLE3 or DFlash is supervised…

Computation and Language · Computer Science 2026-05-29 Haodi Lei , Yafy Li , Haoran Zhang , Shunkai Zhang , Qianjia Cheng , Xiaoye Qu , Ganqu Cui , Bowen Zhou , Ning Ding , Yun Luo , Yu Cheng

The long-output context generation of large reasoning models enables extended chain of thought (CoT) but also drives rapid growth of the key-value (KV) cache, quickly overwhelming GPU memory. To address this challenge, we propose ThinKV, a…

Machine Learning · Computer Science 2026-05-11 Akshat Ramachandran , Marina Neseem , Charbel Sakr , Rangharajan Venkatesan , Brucek Khailany , Tushar Krishna

Speculative decoding is a pivotal technique to accelerate the inference of large language models (LLMs) by employing a smaller draft model to predict the target model's outputs. However, its efficacy can be limited due to the low predictive…

Artificial Intelligence · Computer Science 2024-06-11 Xiaoxuan Liu , Lanxiang Hu , Peter Bailis , Alvin Cheung , Zhijie Deng , Ion Stoica , Hao Zhang

While large language models (LLMs) excel at handling long-context sequences, they require substantial prefill computation and key-value (KV) cache, which can heavily burden computational efficiency and memory usage in both prefill and…

Machine Learning · Computer Science 2026-04-21 Dongwon Jo , Jiwon Song , Yulhwa Kim , Jae-Joon Kim

Memory consumption of the Key-Value (KV) cache represents a major bottleneck for efficient large language model inference. While attention-score-based KV cache pruning shows promise, it faces critical practical limitations: attention scores…

Artificial Intelligence · Computer Science 2025-10-02 Alessio Devoto , Maximilian Jeblick , Simon Jégou

How to efficiently serve LLMs in practice has become exceptionally challenging due to their prohibitive memory and computation requirements. In this study, we investigate optimizing the KV cache, whose memory footprint poses a critical…

Computation and Language · Computer Science 2025-06-10 Akshat Sharma , Hangliang Ding , Jianping Li , Neel Dani , Minjia Zhang

In large language models performing long-form reasoning, the KV cache grows rapidly with decode length, creating bottlenecks in memory and inference stability. Existing reasoning-oriented KV compression has mostly followed an…

Artificial Intelligence · Computer Science 2026-04-14 Buseong Kim , Heejun Gwon

Generative inference in Large Language Models (LLMs) is impeded by the growing memory demands of Key-Value (KV) cache, especially for longer sequences. Traditional KV cache eviction strategies, which discard less critical KV pairs based on…

Computation and Language · Computer Science 2025-03-14 Zhongwei Wan , Xinjian Wu , Yu Zhang , Yi Xin , Chaofan Tao , Zhihong Zhu , Xin Wang , Siqi Luo , Jing Xiong , Longyue Wang , Mi Zhang

Large Language Models (LLMs) rely heavily on Key-Value (KV) caching to minimize inference latency. However, standard KV caches are context-dependent: reusing a cached document in a new context requires recomputing KV states to account for…

Machine Learning · Computer Science 2026-04-20 Chuangtao Chen , Grace Li Zhang , Xunzhao Yin , Cheng Zhuo , Bing Li , Ulf Schlichtmann

Speculative decoding has become a widely adopted technique for accelerating large language model (LLM) inference by drafting multiple candidate tokens and verifying them with a target model in parallel. Its efficiency, however, critically…

Computation and Language · Computer Science 2026-05-19 Shuoyang Sun , Chang Dai , Hao Fang , Kuofeng Gao , Xinhao Zhong , Yi Sun , Fan Mo , Shu-Tao Xia , Bin Chen

The expanding long-context capabilities of large language models are constrained by a significant memory bottleneck: the key-value (KV) cache required for autoregressive generation. This bottleneck is substantial; for instance, a…

Computation and Language · Computer Science 2026-04-20 Yuxuan Zhu , David H. Yang , Mohammad Mohammadi Amiri , Keerthiram Murugesan , Tejaswini Pedapati , Pin-Yu Chen

Speculative sampling reduces the latency of autoregressive decoding for target model LLMs without sacrificing inference quality, by using a cheap draft model to suggest a candidate token and a verification criterion to accept or resample…

Machine Learning · Computer Science 2025-11-21 Rahul Krishna Thomas , Arka Pal

Speculative decoding accelerates large language model (LLM) inference by using a small draft model to generate candidate tokens for a larger target model to verify. The efficacy of this technique hinges on the trade-off between the time…

Computation and Language · Computer Science 2026-03-03 Jiebin Zhang , Zhenghan Yu , Liang Wang , Nan Yang , Eugene J. Yu , Zheng Li , Yifan Song , Dawei Zhu , Xingxing Zhang , Furu Wei , Sujian Li

Inference with Multimodal Large Language Models (MLLMs) is slow due to their large-language-model backbone which suffers from memory bandwidth bottleneck and generates tokens auto-regressively. In this paper, we explore the application of…

Computation and Language · Computer Science 2024-04-16 Mukul Gagrani , Raghavv Goel , Wonseok Jeon , Junyoung Park , Mingu Lee , Christopher Lott

Long-context LLMs have enabled numerous downstream applications but also introduced significant challenges related to computational and memory efficiency. To address these challenges, optimizations for long-context inference have been…

Computation and Language · Computer Science 2025-03-12 Yucheng Li , Huiqiang Jiang , Qianhui Wu , Xufang Luo , Surin Ahn , Chengruidong Zhang , Amir H. Abdi , Dongsheng Li , Jianfeng Gao , Yuqing Yang , Lili Qiu

This work studies how to adaptively recompute key-value (KV) caches for diffusion large language models (DLMs) to maximize prediction accuracy while minimizing decoding latency. Prior methods' decoders recompute QKV for all tokens at every…

Computation and Language · Computer Science 2025-12-30 Quan Nguyen-Tri , Mukul Ranjan , Zhiqiang Shen

Large language models (LLMs) exhibit exceptional performance across a wide range of tasks; however, their token-by-token autoregressive generation process significantly hinders inference speed. Speculative decoding presents a promising…

Computation and Language · Computer Science 2025-03-04 Kai Lv , Honglin Guo , Qipeng Guo , Xipeng Qiu

Memory and computation remain core bottlenecks in long-horizon LLM inference due to the quadratic cost of self-attention and the ever-growing key-value (KV) cache. Existing strategies for memory-bounded inference, such as quantization,…

Machine Learning · Computer Science 2026-03-03 Ngoc Bui , Shubham Sharma , Simran Lamba , Saumitra Mishra , Rex Ying

Large Language Models (LLMs) are increasingly expected to operate over long contexts, yet standard softmax attention incurs a KV cache that grows linearly with sequence length, quickly becoming the bottleneck for long context inference. A…

Computation and Language · Computer Science 2026-05-26 Xintong Yang , Hao Gu , Binxing Xu , Lujun Li , Bei Liu , Jiacheng Liu , Qiyuan Zhu , Sirui Han , Yike Guo