The Inside View cover image

Collin Burns On Discovering Latent Knowledge In Language Models Without Supervision

The Inside View

CHAPTER

Logical Consistency Properties for Sentiment

I can imagine these sorts of techniques extending to other demands so I do think I would be interested and work the pill to in that way one thing I think is like kind of impressive um with your paper. for sentiment uh can do something similar like if a statement is mostly positive if I say calling burns looks good if I say calling burns does not look good or something like this. Our method basically works by taking a set of statements negating them and this you can do it in a purely unsupervised way basically by adding not we're just changing the answer from true to false or yes to no or something like that.

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner