Winners of AI Alignment Awards Research Contest — LessWrong