At least one model can today. https://github.com/cactus-compute/cactus-hybrid
Ahh you found the link I couldn't find with a search.
Yes — they claim that it is likely to be able to assess confidence in whether the statement it just made is correct. But it isn't going to be capable of that while it is answering.
Ahh you found the link I couldn't find with a search.
Yes — they claim that it is likely to be able to assess confidence in whether the statement it just made is correct. But it isn't going to be capable of that while it is answering.