Picture for Dylan Zinsley

Dylan Zinsley

Simple linear attention language models balance the recall-throughput tradeoff

Add code
Feb 28, 2024
Viaarxiv icon