Tags, code fences, or labeled sections make the intended structure easier for the model to infer. Delimiters improve clarity but do not make untrusted content safe or prevent prompt injection by themselves.
A prompt delimiter marks boundaries between instructions, examples, user data, and retrieved content.
Tags, code fences, or labeled sections make the intended structure easier for the model to infer. Delimiters improve clarity but do not make untrusted content safe or prevent prompt injection by themselves.