Word and sentence tokenization explained