llambada

The Chinchilla paper only addresses the contrived use case of a model that is trained once and never used for inference. Since most of the real world compute cost will be in inference, Chinchilla seems to offer little practical guidance.

I don’t think the switch from “you” to “Copilot” is a hallucination OR a mistake. I think this shows that there are two systems being used together: the original Copilot model, and the chat agent. The chat agent is being given instructions for its own behavior, but I suspect it is also evaluating and incorporating the output from the Copilot model (also stuffed into the prompt).