Text this: Beyond Text Generation: Assessing Large Language Models’ Ability to Reason Logically and Follow Strict Rules