What's so special about BERT's layers? A closer look at the NLP pipeline in monolingual and multilingual models.
Wietse de Vries, Andreas van Cranenburgh, Malvina Nissim
Browse the full EMNLP paper archive.
Wietse de Vries, Andreas van Cranenburgh, Malvina Nissim
Browse the full EMNLP paper archive.