Nathanaël Jacquier@nathanaël-jacquier · 1 worksWorks on sparse autoencoders for interpreting vision foundation models.