Annapurna comments on Tracing the Thoughts of a Large Language Model