1 paper
Aina Garí Soler, Matthieu Labeau, Chloé Clavel
When deriving contextualized word representations from language models, a decision needs to be made on how to obtain one for out-of-vocabulary (OOV) words that are segmented into s…