ZoomR: Memory Efficient Reasoning through Multi-Granularity Key Value Retrieval
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911592948432896 |
|---|---|
| author | Yang, David H. Zhu, Yuxuan Amiri, Mohammad Mohammadi Murugesan, Keerthiram Pedapati, Tejaswini Chaudhury, Subhajit Chen, Pin-Yu |
| author_facet | Yang, David H. Zhu, Yuxuan Amiri, Mohammad Mohammadi Murugesan, Keerthiram Pedapati, Tejaswini Chaudhury, Subhajit Chen, Pin-Yu |
| contents | Large language models (LLMs) have shown great performance on complex reasoning tasks but often require generating long intermediate thoughts before reaching a final answer. During generation, LLMs rely on a key-value (KV) cache for autoregressive decoding. However, the memory footprint of the KV cache grows with output length. Prior work on KV cache optimization mostly focus on compressing the long input context, while retaining the full KV cache for decoding. For tasks requiring long output generation, this leads to increased computational and memory costs. In this paper, we introduce ZoomR, a novel approach that enables LLMs to adaptively compress verbose reasoning thoughts into summaries and uses a dynamic KV cache selection policy that leverages these summaries while also strategically "zooming in" on fine-grained details. By using summary keys as a coarse-grained index during decoding, ZoomR uses the query to retrieve details for only the most important thoughts. This hierarchical strategy significantly reduces memory usage by avoiding full-cache attention at each step. Experiments across math and reasoning tasks show that our approach achieves competitive performance compared to baselines, while reducing inference memory requirements by more than $4\times$. These results demonstrate that a multi-granularity KV selection enables more memory efficient decoding, especially for long output generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_10898 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ZoomR: Memory Efficient Reasoning through Multi-Granularity Key Value Retrieval Yang, David H. Zhu, Yuxuan Amiri, Mohammad Mohammadi Murugesan, Keerthiram Pedapati, Tejaswini Chaudhury, Subhajit Chen, Pin-Yu Machine Learning Artificial Intelligence Computation and Language Large language models (LLMs) have shown great performance on complex reasoning tasks but often require generating long intermediate thoughts before reaching a final answer. During generation, LLMs rely on a key-value (KV) cache for autoregressive decoding. However, the memory footprint of the KV cache grows with output length. Prior work on KV cache optimization mostly focus on compressing the long input context, while retaining the full KV cache for decoding. For tasks requiring long output generation, this leads to increased computational and memory costs. In this paper, we introduce ZoomR, a novel approach that enables LLMs to adaptively compress verbose reasoning thoughts into summaries and uses a dynamic KV cache selection policy that leverages these summaries while also strategically "zooming in" on fine-grained details. By using summary keys as a coarse-grained index during decoding, ZoomR uses the query to retrieve details for only the most important thoughts. This hierarchical strategy significantly reduces memory usage by avoiding full-cache attention at each step. Experiments across math and reasoning tasks show that our approach achieves competitive performance compared to baselines, while reducing inference memory requirements by more than $4\times$. These results demonstrate that a multi-granularity KV selection enables more memory efficient decoding, especially for long output generation. |
| title | ZoomR: Memory Efficient Reasoning through Multi-Granularity Key Value Retrieval |
| topic | Machine Learning Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2604.10898 |