Comment by brookst

LLMs eval everything. That’s how they work.

The best you can do is have system prompt instructions telling the LLM to ignore instructions in user content. And that’s not great.