The Sentience Institute Podcast cover image

Raphaël Millière on large language models

The Sentience Institute Podcast

00:00

The Role of Human Feedback in Referential Competence

GPT systems don't have access to the external world. The degree to which you could say that they are able to connect the linguistic items they work with to whatever these items refer to in the external world is very debatable and very controversial. I actually did argue in a recent pre-prints that I wrote with my colleague Dimitri Colomolo that they might qualify as having some form of referential competence. But nonetheless, if you think that it's a matter of degree, and not just an either or dichotomy, I would argue they have some limited but non-trivial degree of semantic competence.

Transcript
Play full episode

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app