Okay, that position seems to me equivalent to the position that models will be more capable at evading monitors when they intrinsically want to evade monitors, than when they are explicitly prompted to evade monitors. Do you also endorse that position? (If not, why isn’t it implied by your stance?)
That position seems pretty unlikely to me (for current models). Models are trained really hard to follow instructions in the prompt to the best of their ability.
Okay, that position seems to me equivalent to the position that models will be more capable at evading monitors when they intrinsically want to evade monitors, than when they are explicitly prompted to evade monitors. Do you also endorse that position? (If not, why isn’t it implied by your stance?)
That position seems pretty unlikely to me (for current models). Models are trained really hard to follow instructions in the prompt to the best of their ability.