English
Related papers

Related papers: Cache What Lasts: Token Retention for Memory-Bound…

200 papers

While Key-Value (KV) cache succeeds in reducing redundant computations in auto-regressive models, it introduces significant memory overhead, limiting its practical deployment in long-sequence scenarios. Existing KV retrieval methods…

Machine Learning · Computer Science 2025-10-14 Wenbo Wu , Qingyi Si , Xiurui Pan , Ye Wang , Jie Zhang

Large language models (LLMs) have demonstrated remarkable performance, but their long-context reasoning remains constrained by the excessive memory required for the Key-Value (KV) cache. This makes KV cache compression a critical step…

Machine Learning · Computer Science 2025-09-30 Xianglong Yan , Zhiteng Li , Tianao Zhang , Haotong Qin , Linghe Kong , Yulun Zhang , Xiaokang Yang

As the field of Large Language Models (LLMs) continues to evolve, the context length in inference is steadily growing. Key-Value Cache (KVCache), the intermediate representations of tokens within LLM inference, has now become the primary…

Computation and Language · Computer Science 2025-04-01 Hailin Zhang , Xiaodong Ji , Yilin Chen , Fangcheng Fu , Xupeng Miao , Xiaonan Nie , Weipeng Chen , Bin Cui

Long-running agentic tasks, such as deep research, require multi-hop reasoning over information distributed across multiple webpages and documents. In such tasks, the LLM context is dominated by tokens from external retrieval, causing…

Artificial Intelligence · Computer Science 2026-03-03 Sanjay Kariyappa , G. Edward Suh

Key-value (KV) caching is widely used to accelerate transformer inference, but its memory cost grows linearly with input length, limiting long-context deployment. Existing token eviction methods reduce memory by discarding less important…

Machine Learning · Computer Science 2026-03-24 Ruijie Miao , Zhiming Wang , Wang Li , Shiwei Wu , Shufan Liu , Yanbing Jiang , Tong Yang

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

Large Language Models (LLMs) have significantly advanced the field of Artificial Intelligence. However, their deployment is resource-intensive, not only due to the large number of model parameters but also because the (Key-Value) KV cache…

Computation and Language · Computer Science 2025-06-05 Yifeng Gu , Zicong Jiang , Jianxiu Jin , Kailing Guo , Ziyang Zhang , Xiangmin Xu

While Large Language Models (LLMs) can theoretically support extensive context windows, their actual deployment is constrained by the linear growth of Key-Value (KV) cache memory. Prevailing compression strategies mitigate this through…

Artificial Intelligence · Computer Science 2026-02-03 Aryan Sood , Tanvi Sharma , Vansh Agrawal

As Large Language Models (LLMs) scale in size and context length, the memory requirements of the key value (KV) cache have emerged as a major bottleneck during autoregressive decoding. The KV cache grows with sequence length and embedding…

Machine Learning · Computer Science 2025-12-09 Sourjya Roy , Shrihari Sridharan , Surya Selvam , Anand Raghunathan

In this study, we investigate whether attention-based information flow inside large language models (LLMs) is aggregated through noticeable patterns for long context processing. Our observations reveal that LLMs aggregate information…

Computation and Language · Computer Science 2025-05-16 Zefan Cai , Yichi Zhang , Bofei Gao , Yuliang Liu , Yucheng Li , Tianyu Liu , Keming Lu , Wayne Xiong , Yue Dong , Junjie Hu , Wen Xiao

The high inference cost of Large Language Models (LLMs) poses challenges, especially for tasks requiring lengthy outputs. However, natural language often contains redundancy, which presents an opportunity for optimization. We have observed…

Computation and Language · Computer Science 2025-11-25 Alfredo Garrachón Ruiz , Tomás de la Rosa , Daniel Borrajo

Large language models (LLMs) have shown strong performance across diverse tasks, but their inference with long input contexts is bottlenecked by memory size and bandwidth. The Key-Value (KV) cache size grows linearly with sequence length…

Machine Learning · Computer Science 2026-05-12 Junkai Zhang , Hang Guo , Luca Benini , Yawei Li

Large language models (LLMs) rely on key-value cache (KV cache) to accelerate decoding by reducing redundant computations. However, the KV cache memory usage grows substantially with longer text sequences, posing challenges for efficient…

Computation and Language · Computer Science 2025-11-18 Yixuan Wang , Shiyu Ji , Yijun Liu , Yuzhuang Xu , Yang Xu , Qingfu Zhu , Wanxiang Che

LLM inference for enterprise applications, such as summarization, RAG, and code-generation, typically observe much longer prompt than generations, leading to high prefill cost and response latency. We present SwiftKV, a novel model…

Machine Learning · Computer Science 2025-06-03 Aurick Qiao , Zhewei Yao , Samyam Rajbhandari , Yuxiong He

Language models handle increasingly long contexts for tasks such as book summarization, but this leads to growing memory costs for the key-value (KV) cache. Many prior works have proposed ways of discarding KVs from memory, but their…

Computation and Language · Computer Science 2025-06-23 Adithya Bhaskar , Alexander Wettig , Tianyu Gao , Yihe Dong , Danqi Chen

The linear growth of the Key-Value (KV) cache is a critical bottleneck in long-form LLM inference. Existing KV compression methods mitigate this by evicting tokens based on importance scores. However, we show that their reliance on global…

Machine Learning · Computer Science 2026-05-25 Junzhe Yang , Xiaoyu Shen

Deploying Vision-Language Models (VLMs) under aggressive low-bit inference remains challenging because inference cost is dominated by the long visual-token prefix during prefill and the growing KV cache during autoregressive decoding. Token…

Computer Vision and Pattern Recognition · Computer Science 2026-04-21 Xinqing Li , Xin He , Xindong Zhang , Ming-Ming Cheng , Lei Zhang , Yun Liu

Long-context LLM serving is bottlenecked by the cost of attending over ever-growing KV caches. Dynamic sparse attention promises relief by accessing only a small, query-dependent subset of the KV state per decoding step and extending the KV…

Machine Learning · Computer Science 2026-04-30 Zihan Zhao , Baotong Lu , Shengjie Lin , Yizou Chen , Jing Liu , Yanqi Zhang , Ziming Miao , Ming-Chang Yang , Haiying Shen , Qi Chen , Fan Yang

Large language models (LLMs) with extended context windows enable powerful downstream applications but impose significant memory overhead, as caching all key-value (KV) states scales linearly with sequence length and batch size. Existing…

Computation and Language · Computer Science 2025-10-10 Yuzhe Gu , Xiyu Liang , Jiaojiao Zhao , Enmao Diao

In Text-to-SQL tasks, existing LLM-based methods often include extensive database schemas in prompts, leading to long context lengths and increased prefilling latency. While user queries typically focus on recurrent table sets-offering an…

Computation and Language · Computer Science 2026-01-14 Jinbo Su , Yuxuan Hu , Cuiping Li , Hong Chen , Jia Li , Lintao Ma , Jing Zhang