Writing down intermediate calculations is not a concept we invented. In our paper we call this 'elicitive prompting' and mention chain-of-thought prompting and the scratchpad papers as previous examples of this.

​

I'm super excited about elicitive prompts (self-ask is in that category too)! I think they're going to enable us to get much more out of these models.

​

And yes, just like we can integrate Google Search we can also integrate lots of other systems, I'm really excited to see how this research direction develops!