I've always been curious. Are tags like that one: "<system-reminder>" useful at all? Is the LLM training altered to give a special meaning to specific tags when they are found?
Can a user just write those magic tags (if they knew what they are) and alter the behavior of the LLM in a similar manner?
You can just make them up, and ask it to respond with specific tags, too.
Like “Please respond with the name in <name>…</name> tags and the <surname>.”
It’s one of the approaches to forcing structured responses, or making it role-play multiple actors in one response (having each role in its tags), or asking it to do a round of self-critique in <critique> tags before the final response, etc.
- Circuit breakers when it seem like it's stuck in a loop
- Warnings about running low on context
- Reminders about task lists (or anything)
- All sorts of warnings about whateverOkay, I know I shouldn't anthropomorphize, but I couldn't prevent myself from thinking that this was a bit of a harsh way of saying things :(