Saved in:
Bibliographic Details
Main Authors: Berestizshevsky, Konstantin, Andri, Renzo, Cavigelli, Lukas
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.08363
Tags: Add Tag
No Tags, Be the first to tag this record!
Table of Contents:
  • We present Top-Theta (Top-$θ$) Attention, a training-free method for sparsifying transformer attention during inference. Our key insight is that static, per-head thresholds can be calibrated to retain the desired constant number of significant elements per attention row. This approach enables content-based sparsity without retraining, and it remains robust across data domains. We further introduce compensation techniques to preserve accuracy under aggressive sparsification, establishing attention thresholding as a practical and principled alternative to top-k attention. We provide extensive evaluation on natural language processing tasks, showing that Top-$θ$ achieves 3-10x reduction in V-cache usage and up to 10x fewer attention elements during inference while degrading no more than 1% in accuracy.