139 - Coherent Long Story Generation, with Kevin Yang
NLP Highlights
00:00
How to Generate Stories From GBDP Models
The problems that we deal with on the scale of 2000 words are just like very different problems with people dealt with before when they were working with 100 or 200 words. So our approach is trying to use GBP3 as a base model that is pretty good at generating locally fluent paragraphs but trying to build the framework for prompting that goes around it. We're actually in a sense looking at a new research problem because the kinds of problems we have to deal with regarding generating content from GBDP are very different from what people looked at earlier.
Transcript
Play full episode
