The Devil is in the Condition Numbers: Why is GLU Better than non-GLU Structure? 事件

The Devil is in the Condition Numbers: Why is GLU Better than non-GLU Structure? arXiv:2605.20749v2 Announce Type: replace-cross Abstract: Gated Linear Units (GLU) and their variants are widely adopted in modern open-source large language model architectures and consistently outperform their non-gated counterparts, yet the underlying reasons for this advantage remain unclear. In this work, we study GLU by analyzing two-layer networks in the neural tangent kernel (NTK) regime. Our analysis revea