Explorations into the recently proposed Taylor Series Linear Attention
Implementation of Agent Attention in Pytorch
LEAP: Linear Explainable Attention in Parallel for causal language modeling with O(1) path length, and O(1) inference
