{"type":"link","version":"1.0","title":"Context-extension methods that sparsify or drop attention systematically lose long-context accuracy versus methods that keep full attention and only rescale positions","author_name":"AI Archs","author_url":"https://ai-arch.pages.dev","provider_name":"AI Archs","provider_url":"https://ai-arch.pages.dev","url":"https://ai-arch.pages.dev/n/approximate-attention-trades-long-context-accuracy-for-cheaper-compute","thumbnail_url":"https://ai-arch.pages.dev/android-chrome-512x512.png","thumbnail_width":512,"thumbnail_height":512}