Papers
arxiv:2508.02124

Trainable Dynamic Mask Sparse Attention

Published on Aug 4
· Submitted by JingzeShi on Aug 4
Authors:
,
,
,
,
,
,

Abstract

A dynamic mask sparse attention mechanism, DMA, improves long-context modeling in large language models by reducing computational complexity while maintaining information fidelity.

AI-generated summary

In large language models, the demand for modeling long contexts is constantly increasing, but the quadratic complexity of the standard self-attention mechanism often becomes a bottleneck. Although existing sparse attention mechanisms have improved efficiency, they may still encounter issues such as static patterns or information loss. We introduce a trainable dynamic mask sparse attention mechanism, Dynamic Mask Attention, which effectively utilizes content-aware and position-aware sparsity. DMA achieves this through two key innovations: First, it dynamically generates content-aware sparse masks from value representations, enabling the model to identify and focus on critical information adaptively. Second, it implements position-aware sparse attention computation that effectively skips unnecessary calculation regions. This dual-sparsity design allows the model to significantly reduce the computational complexity of important information while retaining complete information, achieving an excellent balance between information fidelity and computational efficiency. We have verified the performance of DMA through comprehensive experiments. Comparative studies show that DMA outperforms multi-head attention, sliding window attention, multi-head latent attention, and native sparse attention in terms of perplexity under Chinchilla Scaling Law settings. Moreover, in challenging multi-query associative recall tasks, DMA also demonstrates superior performance and efficiency compared to these methods. Crucially, in the evaluation of a 1.7B parameter model, DMA significantly outperforms multi-head attention in both standard benchmark performance and the challenging needle-in-a-haystack task. These experimental results highlight its capability to balance model efficiency and long-context modeling ability effectively.

Community

Paper submitter

LGTM

Sign up or log in to comment

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2508.02124 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2508.02124 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2508.02124 in a Space README.md to link it from this page.

Collections including this paper 2