chunked/nltk_chunking/model_doc_herbert/chunk_2.txt

The abstract from the paper is the following:
In recent years, a series of Transformer-based models unlocked major improvements in general natural language
understanding (NLU) tasks.