It’s not best fit or regression lines since different model sizes/capabilities are on the graph. Why would you best fit to the average of the new 8B, 70B and 405B llama 3.1 models? These are just rough trends lines of the frontier

→ View original post on X — @thom_wolf,