Don't Share Information Exfohazardous on Others' AI-Risk Models — AI Alignment Forum