PIGuard: Prompt Injection Guardrail via Mitigating Overdefense for Free(injecguard.github.io)

15 ポイント投稿者 mettamage3 か月前5 コメント

injecguard.github.io

PIGuard: Prompt Injection Guardrail via Mitigating Overdefense for Free

https://injecguard.github.io/

6 コメント

I was playing around with some prompt injection guard rails frameworks. I know they don't mitigate attack classes, but they at least do something. I just got a bit miffed about the high false positive rates I saw in my own testing.

This one has a low false positive rate. And I thought that was interesting.

ekns3 か月前

There is a simple way to mitigate prompt injection. Just check metadata only: is this action by the LLM suspicious given trusted metadata, blanking out the data

carterschonwald3 か月前

while i cant speak regarding arbitrary prompt injections, ive been using a simple approach i add to any llm harness i use, that seems to solve turn or role confusion being remotely viable.

i really need to test my toolkit (carterkit) augmented harnesses on some of the more respectavle benchmarks

ninju3 か月前

You misspelled 'execute' in the video ;)

bastawhiz3 か月前

And it still did the right thing. Which I think makes the demo slightly more impressive.

JoshBlythe3 か月前