LessWrong (30+ Karma) - “Twitter thread on AI safety evals” by Richard_Ngo
Sign in to continue reading, translating and more.