X · @teortaxesTex
· X / Twitter
GDP's hot take: Compressed Sparse Attention was not a rejection of MLA (DSA ver.) but its dialectical rebirth. Similarities go pretty deep. MLA-NSA(st…
GDP's hot take: Compressed Sparse Attention was not a rejection of MLA (DSA ver.) but its dialectical rebirth. Similarities go pretty deep. MLA-NSA(stillborn)-DSA-CSA is one evolution of the core compression ideology.(+ awesome diagram reposted by @ZhihuFrontier)GDP: @tugot17 If you see closely CSA is MLA++. Look at th