The Nonlinear Library - LW - Twitter thread on AI safety evals by Richard Ngo
Sign in to continue reading, translating and more.
