1 paper
Gokul Srinivasagan, Kai Hartung, Munir Georges
Masked language modeling has become a standard pretraining objective for training encoder-based language models. In this approach, certain tokens in the input are masked, and the m…