Local Linear Attention: An Optimal Interpolation of Linear and Softmax Attention For Test-Time Regression
机构 * Snowflake AI Research(Snowflake AI研究机构) ; Northwestern University(西北大学) ; University of Washington(华盛顿大学)
大厂专区
机构 * Snowflake AI Research(Snowflake AI研究机构) ; Northwestern University(西北大学) ; University of Washington(华盛顿大学)