Text this: The Use of Large Language Models for Marking Datasets in Natural Language Processing