When you generate a text with an LLM, you always have some choice. So you can sample in a way that is very likely under your watermark scheme, and unlikely otherwise
but that's what I'm saying, When you ask for exclamation marks after each word that must change the likelihoods of next token by quite a bit. You then remove the marks which hides the fact that you just changed every word without loss of meaning
When you generate a text with an LLM, you always have some choice. So you can sample in a way that is very likely under your watermark scheme, and unlikely otherwise