Language Models Can Control Their Own Attention
Language models spend most of their attention on a small fraction of context, yet they read the entire KV cache to find the few tokens that matter. If the user asks about a previous detail in a 1M-token conversation, global attention layers must scan the full context to generate each token of the reply. A prominent approach mitigates this cost by pre-selecting relevant tokens via lightweight…
We haven't written up this one. arXiv cs.AI has the full story — the link below goes straight to it.
