That was my first thought too -- instead of talk like a caveman you could turn off reasoning, with probably better results.
Additionally, LLMs do not actually operate in text; much of the thinking happens in a much higher dimensional space that just happens to be decoded as text.
So unless the LLM was trained otherwise, making it talk like a caveman is more than just theoretically turning it into a caveman.
Additionally, LLMs do not actually operate in text; much of the thinking happens in a much higher dimensional space that just happens to be decoded as text.
So unless the LLM was trained otherwise, making it talk like a caveman is more than just theoretically turning it into a caveman.