(experimental) RL agents should be more aligned if they do not know whether they are in simulation or in the real world
By chatting or signing in you agree to the Terms and chat-message logging (revocable in History).