hard20 pts · first answer countsllm-safetyjailbreaksguardrails
Detecting gibberish suffixes
A defence computes the perplexity of incoming prompts and flags those with unusually high values. Which attack is this best at catching, and what is its weakness?
Discussion 0
Join free to post in the discussion.
Related questions
- Jailbreak or prompt injectionPrompt Engineer
- Customer data in promptsPrompt Engineer
- Top risk in the OWASP LLM listPrompt Engineer
- What "guardrails" meansPrompt Engineer
- Checking before and after the modelPrompt Engineer
- Tool access for an FAQ botPrompt Engineer
- Calling the LLM from the browserPrompt Engineer
- One user sending 50,000 requestsPrompt Engineer