Some discussion from DeepMind researchers about potential for language models to train on data containing outputs of other LMs. From this paper: https://t.co/XLMl6AL6lS https://t.co/jYvF2txueZ pic.twitter.com/m2CEOVUG7R
— Jack Clark (@jackclarkSF) March 30, 2021
from Twitter https://twitter.com/jackclarkSF
March 30, 2021 at 01:20PM
via IFTTT
No comments:
Post a Comment