Comment by tmnvdb

This is a hard question for language models since it targets one of their known weaknesses.