Rectified Sparse Attention for Efficient Long-Sequence Generation

ACL ARR 2026 January Submission552 Authors

Keywords: Sparse Attention, Test Time Scaling, Long Generation

Abstract: Efficient long-sequence generation is a critical challenge for Large Language Models. While recent sparse decoding methods improve efficiency, they suffer from KV cache misalignment, where approximation errors accumulate and degrade generation quality. In this work, we propose Rectified Sparse Attention (ReSA), a simple yet effective method that combines block-sparse attention with periodic dense rectification. By refreshing the KV cache at fixed intervals using a dense forward pass, ReSA bounds error accumulation and preserves alignment with the pretraining distribution. Experiments across math reasoning, language modeling, and retrieval tasks demonstrate that ReSA achieves near-lossless generation quality with significantly improved efficiency. Notably, ReSA delivers up to 3.77x end-to-end speedup under decoding at 256K sequence length, making it a practical solution for scalable long-context inference.

Paper Type: Long

Research Area: Language Models

Research Area Keywords: sparse models

Contribution Types: NLP engineering experiment, Approaches low compute settings-efficiency

Languages Studied: English

Submission Number: 552