Summary. This interactive notebook demonstrates how to train a reinforcement-learning agent in text-based environments using an LLM-parameterized policy. The ability to understand and generate natural ...