[joshspankit]

It seems like the right solution is around the corner: placeholders for these kinds of strings (uuid, hash, etc)

Why should an LLM even have these types of IDs anywhere in the prediction pipeline?