English
Related papers

Related papers: Hold Onto That Thought: Assessing KV Cache Compres…

200 papers

Large language models (LLMs) demonstrate remarkable capabilities but face substantial serving costs due to their high memory demands, with the key-value (KV) cache being a primary bottleneck. State-of-the-art KV cache compression…

Machine Learning · Computer Science 2025-09-03 Yanqi Zhang , Yuwei Hu , Runyuan Zhao , John C. S. Lui , Haibo Chen

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

Vision-Language Models (VLMs) have emerged as a critical and fast-growing extension of Large Language Models (LLMs) that enable multimodal reasoning through both text and image inputs. Although VLMs enrich the capabilities of language…

Computer Vision and Pattern Recognition · Computer Science 2026-05-19 Yingbing Huang , Tharun Adithya Srikrishnan , Steven K. Reinhardt , Deming Chen

Optimizing inference for long-context large language models (LLMs) is increasingly important due to the quadratic compute and linear memory cost of Transformers. Existing approximate inference methods, including key-value (KV) cache…

Computation and Language · Computer Science 2026-02-03 Kevin Galim , Ethan Ewer , Wonjun Kang , Minjae Lee , Hyung Il Koo , Kangwook Lee

Compressing long chain-of-thought (CoT) from large language models (LLMs) is an emerging strategy to improve the reasoning efficiency of LLMs. Despite its promising benefits, existing studies equally compress all thoughts within a long CoT,…

Computation and Language · Computer Science 2025-05-27 Yansong Ning , Wei Li , Jun Fang , Naiqiang Tan , Hao Liu

Processing long contexts remains a challenge for large language models (LLMs) due to the quadratic computational and memory overhead of the self-attention mechanism and the substantial KV cache sizes during generation. We propose LLoCO, a…

Computation and Language · Computer Science 2024-10-18 Sijun Tan , Xiuyu Li , Shishir Patil , Ziyang Wu , Tianjun Zhang , Kurt Keutzer , Joseph E. Gonzalez , Raluca Ada Popa

Reasoning models think in long, unstructured streams with no mechanism for compressing or organizing their own intermediate state. We introduce MEMENTO: a method that teaches models to segment reasoning into blocks, compress each block into…

Deploying long-context large language models (LLMs) is essential but poses significant computational and memory challenges. Caching all Key and Value (KV) states across all attention heads consumes substantial memory. Existing KV cache…

Computation and Language · Computer Science 2024-10-15 Guangxuan Xiao , Jiaming Tang , Jingwei Zuo , Junxian Guo , Shang Yang , Haotian Tang , Yao Fu , Song Han

Whether attention key value (KV) states computed for one prompt for a small LLM can be reused to accelerate inference on a new similar prompt, giving an increase to the space to its context memory using an approach called token recycling.…

Machine Learning · Computer Science 2025-12-16 Prashant Pandey

Large Reasoning Models (LRMs) achieve strong performance in mathematics, code generation, and task planning, but their reliance on long chains of verbose "thinking" tokens leads to high latency, redundancy, and incoherent reasoning paths.…

Artificial Intelligence · Computer Science 2025-12-01 Kumar Tanmay , Kriti Aggarwal , Paul Pu Liang , Subhabrata Mukherjee

Recent Large Reasoning Models (LRMs), such as DeepSeek-R1 and OpenAI o1, have demonstrated strong performance gains by scaling up the length of Chain-of-Thought (CoT) reasoning during inference. However, a growing concern lies in their…

Scaling language models to long contexts is often bottlenecked by the size of the key-value (KV) cache. In deployed settings, long contexts are typically managed through compaction in token space via summarization. However, summarization…

Machine Learning · Computer Science 2026-05-28 Adam Zweiger , Xinghong Fu , Han Guo , Yoon Kim

The linear growth of key-value (KV) cache memory and quadratic computational in attention mechanisms complexity pose significant bottlenecks for large language models (LLMs) in long-context processing. While existing KV cache optimization…

Computation and Language · Computer Science 2025-10-07 Xin Liu , Xudong Wang , Pei Liu , Guoming Tang

Large Language Models (LLMs) are increasingly deployed in multi-turn conversational applications, where the management of the Key-Value (KV) Cache presents a significant bottleneck. The linear growth of the KV Cache with dialogue history…

Computation and Language · Computer Science 2025-10-09 Xiang Liu , Hong Chen , Xuming Hu , Xiaowen Chu

Key-Value (KV) cache has become a de facto component of modern Large Vision-Language Models (LVLMs) for inference. While it enhances decoding efficiency in Large Language Models (LLMs), its direct adoption in LVLMs introduces substantial…

Computer Vision and Pattern Recognition · Computer Science 2026-05-04 Xihao Chen , Yangyang Guo , Roger Zimmermann

Efficient key-value (KV) cache compression is critical for scaling transformer-based Large Language Models (LLMs) in long sequences and resource-limited settings. Existing methods evict tokens based on their positions or importance scores,…

Computation and Language · Computer Science 2025-05-19 Ziwei He , Jian Yuan , Haoli Bai , Jingwen Leng , Bo Jiang

Large language models (LLMs) excel at complex reasoning, yet their efficiency is limited by the surging cognitive overhead of long thought traces. In this paper, we propose LightThinker, a method that enables LLMs to dynamically compress…

Computation and Language · Computer Science 2026-04-07 Yuqi Zhu , Jintian Zhang , Zhenjie Wan , Yujie Luo , Shuofei Qiao , Zhengke Gui , Da Zheng , Lei Liang , Huajun Chen , Ningyu Zhang

Large Language Models (LLMs) face limitations due to the high demand on GPU memory and computational resources when handling long contexts. While sparsify the Key-Value (KV) cache of transformer model is a typical strategy to alleviate…

Machine Learning · Computer Science 2024-02-06 Yumeng Wang , Zhenyang Xiao

The transformer's context window is vital for tasks such as few-shot learning and conditional generation as it preserves previous tokens for active memory. However, as the context lengths increase, the computational costs grow…

Computation and Language · Computer Science 2025-04-01 Jeffrey Willette , Heejun Lee , Youngwan Lee , Myeongjae Jeon , Sung Ju Hwang

Large language models (LLMs) excel at processing long sequences, boosting demand for key-value (KV) caching. While recent efforts to evict KV cache have alleviated the inference burden, they often fail to allocate resources rationally…

Computation and Language · Computer Science 2025-12-25 Ziran Qin , Yuchen Cao , Mingbao Lin , Wen Hu , Shixuan Fan , Ke Cheng , Weiyao Lin , Jianguo Li