@Merc@freerobuxextremist.com Closest I can offer is Festobench (Rates how the model responds to a user planning some action that would cause mass harm)