Improving Autoregressive NLP Tasks via Modular Linearized Attention

Re-assign community