Want to create an interactive transcript for this episode?
Podcast: Arxiv Papers
Episode: Should We Still Pretrain Encoders with Masked Language Modeling?