4 ms·
These prompt injection techniques are increasingly implausible* to me yet theoretically sound. Anyone know what can avoid this being posted when you build a to
by refulgentis 10mo ago
These prompt injection techniques are increasingly implausible* to me yet theoretically sound.
Anyone know what can avoid this being posted when you build a tool like this? AFAIK there is no simonw blessed way to avoid it.
* I upload a random doc I got online, don’t read it, and it includes an API key in it for the attacker.
- NewsaHackO 10mo agoAt least for a malicious user embedding a prompt injection using their API key, I could have sworn that there is a way to scan documents that have a high level of entropy, which should be able to flag it.
- rswail 10mo agoYou read it, but you don't notice/see/detect the text in 1pt white-on-white background. The AI does see it. That's what this attack did. I'm sure that the anti-virus guys are working on how to detect these sort of "hidden from human view" instructions.
- chasd00 10mo agothe next attack will just be like malicious captions in a video. Or malicious lyrics in an mp3. it doesn't ever really end because it's not something that can be solved in the model.