Sign in

Large language models implicitly learn to straighten neural sentence trajectories to construct a predictive representation of natural language.

Eghbal A. HosseiniEvelina Fedorenko
Published in: CoRR (2023)
Keyphrases