# ProxyAttn: Guided Sparse Attention via Representative Heads

arXiv: https://arxiv.org/abs/2509.24745 · HTML: https://mlsanity.com/paper/2509.24745/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 35 of its 40 references that are in the map.
- First posted on arXiv 29 Sep 2025.

## Builds on (most cited first, 30 of 35)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [SWE-bench: Can Language Models Resolve Real-World GitHub Issues?](https://mlsanity.com/paper/2310.06770/) (2023)
- [FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioning](https://mlsanity.com/paper/2307.08691/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [Big Bird: Transformers for Longer Sequences](https://mlsanity.com/paper/2007.14062/) (2020)
- [LongBench: A Bilingual, Multitask Benchmark for Long Context Understanding](https://mlsanity.com/paper/2308.14508/) (2023)
- [GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints](https://mlsanity.com/paper/2305.13245/) (2023)
- [RULER: What's the Real Context Size of Your Long-Context Language Models?](https://mlsanity.com/paper/2404.06654/) (2024)
- [Qwen2.5-Coder Technical Report](https://mlsanity.com/paper/2409.12186/) (2024)
- [DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model](https://mlsanity.com/paper/2405.04434/) (2024)
- [Fast Transformer Decoding: One Write-Head is All You Need](https://mlsanity.com/paper/1911.02150/) (2019)
- [Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention](https://mlsanity.com/paper/2502.11089/) (2025)
- [Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models](https://mlsanity.com/paper/2503.09567/) (2025)
- [MInference 1.0: Accelerating Pre-filling for Long-Context LLMs via Dynamic Sparse Attention](https://mlsanity.com/paper/2407.02490/) (2024)
- [LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks](https://mlsanity.com/paper/2412.15204/) (2024)
- [∞Bench: Extending Long Context Evaluation Beyond 100K Tokens](https://mlsanity.com/paper/2402.13718/) (2024)
- [DuoAttention: Efficient Long-Context LLM Inference with Retrieval and Streaming Heads](https://mlsanity.com/paper/2410.10819/) (2024)
- [MoBA: Mixture of Block Attention for Long-Context LLMs](https://mlsanity.com/paper/2502.13189/) (2025)
- [Qwen2.5-1M Technical Report](https://mlsanity.com/paper/2501.15383/) (2025)
- [XAttention: Block Sparse Attention with Antidiagonal Scoring](https://mlsanity.com/paper/2503.16428/) (2025)
- [SeerAttention: Learning Intrinsic Sparse Attention in Your LLMs](https://mlsanity.com/paper/2410.13276/) (2024)
- [FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference](https://mlsanity.com/paper/2502.20766/) (2025)
- [A Comprehensive Survey on Long Context Language Modeling](https://mlsanity.com/paper/2503.17407/) (2025)
- [Attention Heads of Large Language Models: A Survey](https://mlsanity.com/paper/2409.03752/) (2024)
- [A General Survey on Attention Mechanisms in Deep Learning](https://mlsanity.com/paper/2203.14263/) (2022)
- [MiniCPM4: Ultra-Efficient LLMs on End Devices](https://mlsanity.com/paper/2506.07900/) (2025)

## Built on by (most cited first, 9 of 9)

- [On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability](https://mlsanity.com/paper/2608.30320/) (2026)
- [VideoNSA: Native Sparse Attention Scales Video Understanding](https://mlsanity.com/paper/2510.02295/) (2025)
- [FlashPrefill: Instantaneous Pattern Discovery and Thresholding for Ultra-Fast Long-Context Prefilling](https://mlsanity.com/paper/2603.06199/) (2026)
- [(Sparse) Attention to the Details: Preserving Spectral Fidelity in ML-based Weather Forecasting Models](https://mlsanity.com/paper/2604.16429/) (2026)
- [EchoKV: Efficient KV Cache Compression via Similarity-Based Reconstruction](https://mlsanity.com/paper/2603.22910/) (2026)
- [PIVOT: Efficient Query-Group Indexing for Token-Level Sparse Attention](https://mlsanity.com/paper/2607.24593/) (2026)
- [UniPrefill: Universal Long-Context Prefill Acceleration via Block-wise Dynamic Sparsification](https://mlsanity.com/paper/2605.06221/) (2026)
- [CoSA: Accelerating Long-Context Inference via Proxy-Kernel Co-Designed Sparse Attention](https://mlsanity.com/paper/2607.25291/) (2026)
- [The Sirens' Song: When Proximal Background Context Overshadows Distant Evidence](https://mlsanity.com/paper/2609.26718/) (2026)

## Abstract (arXiv)

The quadratic complexity of attention mechanisms limits the efficiency of Large Language Models (LLMs) on long-text tasks. Recently, methods that dynamically estimate block importance have enabled efficient block sparse attention, leading to significant acceleration in long-text pre-filling of LLMs. However, their coarse-grained estimation inevitably leads to performance degradation at high sparsity rates. In this work, we propose ProxyAttn, a training-free sparse attention algorithm that achieves more precise block estimation by compressing the dimension of attention heads. Based on our observation of the similarity among multiple attention heads, we use the scores of pooled representative heads to approximate the scores for all heads. To account for the varying sparsity among heads, we also propose a block-aware dynamic budget estimation method. By combining the scores from representative proxy heads with multi-head dynamic budgets, we achieve a more fine-grained block importance evaluation at low computational cost. Experiments on a variety of mainstream models and extensive benchmarks confirm the underlying similarity among attention heads. Leveraging a fine-grained estimation, the proposed method achieves substantial gains in performance and efficiency compared to existing methods. More precisely, ProxyAttn can achieve up to 10.3x attention acceleration and 2.4x prefilling acceleration without significant performance loss. Our code is available at https://github.com/wyxstriker/ProxyAttn.