LSG Attention: Extrapolation of pretrained Transformers to long sequences Paper • 2210.15497 • Published Oct 13, 2022 • 1 • 1