
AB
Ambroise Borbely, Mathieu Salzmann, Olivier Verscheure
· 1 min read
ResearcharXiv cs.CL
HiSA: Sparsifying the Hidden Dimension of RoPE Attention in LLMs
arXiv:2611.26363v1 Announce Type: cross
Abstract: Attention in large language models (LLMs) scores every query against every key with a full dot product. We investigate whether this dot product must be computed completely for attention to be effective. We propose Hidden Sparse Attention (HiSA), which keeps the query and key projections intact but only computes a partial dot product, following a static, parameter-free rule set by the token positions: attention becomes sparse in the hidden (feature) dimension, an axis much less explored than token sparsity, rather than over tokens as in almost all sparse attention methods. Based on the structure of rotary position embeddings (RoPE), we decompose keys and queries into parts which can be omitted when computing attention. By training LLMs from scratch with HiSA, we find that this partial dot product does not hurt but improves performance compared with the complete product while computing only about a third of the query-key products. During inference HiSA improves decoding speed and shrinks memory footprint. We also extend HiSA to FlashAttention-3 kernels for training and inference. Finally, we motivate our approach by showing that dense attention learns to suppress the attention score parts that HiSA drops. Code is available at https://github.com/ambroise-b/hisa.
Original source
This story was published by arXiv cs.CL and written by Ambroise Borbely, Mathieu Salzmann, Olivier Verscheure. SyncAI.news shows a preview; the complete article is on the publisher's site.
Read the full story on arxiv.org


