PAP
Multi-Gate Residuals
May 22, 2026
Abstract
While Attention Residuals has shown some effectiveness in addressing the widespread issue of unbounded activation growth across deep residual layers, it inevitably incurs significant communication overhead. To circumvent this bottleneck, we propose Multi-Gate Residuals (MGR), which stabilizes activation scales without additional communication burden. It utilizes a straightforward scoring and gating mechanism to maintain multi-stream context, coupled with Attention Pooling to extract hidden states from the stream states. Empirical experiments demonstrate that MGR is practical for large-scale training and deployment, offering tangible performance improvements over existing architectures.
Select text to highlight · click a highlight to remove · saved in this browser only
Authors
Zhizhan Zheng, Feiyun Zhang, Shuchun Liu, Tian Xia, Xi Liu, Dasheng Hu, Hongquan Zhou
Your notes (browser-local)
savedarXiv:2605.23259