Last released Sep 27, 2022
Linear Explainable Attention in Parallel (LEAP) for causal language modeling (also implements fastformer)
Supported by