Explore two AI experiments investigating contextual token representations, Transformer hidden states, and the accuracy of structured classification outputs.