#artificial intelligence
May 2026
Attention Sinks and Outliers in Attention Residuals
OASIS is proposed, an outlier- and sink-aware method that stabilizes dual-normalized attention-residual architectures through explicit null routing and token-to-depth null coupling and offers insight into the low-bit sensitivity observed in AttnResidual.
Haozheng Luo, Hao-Ran Dai, Shao-Ling Zhang et al.
· arXiv.org · 2 citations