Focused Transformer: Contrastive Training for Context Scaling
作者:Szymon Tworkowski, Konrad Staniszewski, Mikołaj Pacek, Yuhuai Wu, Henryk Michalewski, Piotr Miłoś · 单位:Institute of Mathematics, Polish Academy of Sciences, University of Warsaw, Google DeepMind · 来源:arXiv · 发布日期:2024/03