Fetching the paper…

Improving Pretrained Cross-Lingual Language Models via Self-Labeled Word Alignment · Around