Stanford NLP Group sur Twitter : "Evidence that large pre-trained language models do successfully induce world models—because they enhance the model’s ability to predict the future)."
Tags:
About This Document
File info
Documents with similar tags (experimental)
2019-02-14 About