> possibility of "LLMs can reason not like a human"

what would be the difference between not reasoning and reasoning not like a human?