Agree with this. A few prompt variants: \* What if you allow the model to do Cha...

		Closi 11 months ago \| parent \| context \| favorite \| on: Something weird is happening with LLMs and chess Agree with this. A few prompt variants: * What if you allow the model to do Chain of Thought (explicitly disallowed in this experiment) * What if you explain the board position at each step to the model in the prompt, so it doesn't have to calculate/estimate it internally.

They also tested GPT-o1, which is always CoT. Yet it is still worse.