{"type":"link","version":"1.0","title":"When a block can gate between local and long-term attention, heads polarize into specialists and mixers during training","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/attention-heads-self-segregate-into-local-and-memory-roles","thumbnail_url":"https://ai-arch.pages.dev/android-chrome-512x512.png","thumbnail_width":512,"thumbnail_height":512}