Comment by radicaldreamer

I think the problem with that is that Grok has likely been prompted to do that in the system prompt or some prompts that get added for questions about Elon. That doesn't reflect on the actual reasoning or generalization abilities of the underlying model most likely.