x
Any work on honeypots (to detect treacherous turn attempts)? — AI Alignment Forum