I was under the impression that the purpose of "system" prompt is to encode the instruction boundary explicitly to reduce the risk of injection. Do you enforce

I was under the impression that the purpose of "system" prompt is to encode the instruction boundary explicitly to reduce the risk of injection. Do you enforce some kind of security invariant that we could rely on? For example, does the alignment regiment include adversarial demonstrations so that out-of-order instruction-following (such as contradicting preceding) is penalised?