chunked/nltk_chunking/model_doc_tapex/chunk_6.txt

The abstract from the paper is the following:
Recent progress in language model pre-training has achieved a great success via leveraging large-scale unstructured textual data.