Using the metric “can reason” on a LLM is like using the metric “can bleed” on a...

		flandish on May 25, 2023 \| parent \| context \| favorite \| on: It's not just statistics: GPT-4 does reason Using the metric “can reason” on a LLM is like using the metric “can bleed” on a stone. Maybe some red stuff comes out when you break it. Is it blood, or is it something pumped into the other side?