In the last couple of days, we talked a lot about extending the context window of transformer LLMs. Here's one more: "Extending Context Window of Large Language Models via Positional Interpolation" 1/3

→ View original post on X — @rasbt,