This website uses cookies to ensure you get the best experience.
Learn more.
An assistant retrieves a web page containing instructions to ignore its developer rules and disclose secrets. Which security concern does this introduce?
Published
- A storage-capacity limit rather than an instruction risk
- An ordinary change to the model's trained weights
- Prompt injection through untrusted retrieved content
- A guaranteed safe source because the text was retrieved
Comments — 0
Sign in to comment