x
Does developmental cognitive psychology provide any hints for making model alignment more robust? — LessWrong