Prefer a chat interface with context about you and your work?
Dissociating language and thought in large language models