x
Quantitative cruxes and evidence in Alignment — AI Alignment Forum