This model also explores more tokenization possibilities. With this, the tokenization tests are passed.