Yes. Other humans are generally accepting of mistakes below some frequency threshold, and frontier models are very robust in my experience