Ah yes we have some benchmarks on this sort of misguided prompt trap, so it should perform well on this