The Promise of Language Models for Search: Generative Information Retrieval
Neural Search Talks — Zeta Alpha
00:00
The Future of Open-Ended Question Answering
Andrew Keen: I'm very interested in the whole space of using language models to evaluate all language models. The difficulty with open-ended question answering is that you don't know exactly how the correct answer should be written. There's things like Burt's score where they're trying to use a language model to look at the agreement. And then with large language models, I think everyone knows better than me. But there's more recent stuff like this, I think, constitutional AI. They do something like this, where they're asking a languagemodel to kind of evaluate whether this looks right.
Transcript
Play full episode
