Re:New LLM jailbreak bypasses all major FMs
by BriggyDwiggs42 on Friday April 25, 2025 @06:13PM
(#43798995)
In what world would it respond with bomb building instructions?
[ Parent
| on HN ]
Re:New LLM jailbreak bypasses all major FMs
by QuadmasterXLII on Friday April 25, 2025 @08:16PM
(#43799744)
if it used search and ingested a malicious website, fir example
Re:New LLM jailbreak bypasses all major FMs
by BriggyDwiggs42 on Saturday April 26, 2025 @12:47AM
(#43801023)
Fair, but if it happens upon that in the top search results of an innocuous search, maybe the LLM isn’t the problem.
Re:New LLM jailbreak bypasses all major FMs
by __MatrixMan__ on Friday April 25, 2025 @06:36PM
(#43799124)
If I were to make a list of fun things, I think that blowing stuff up would feature in the top ten. It's not unreasonable that an LLM might agree.
Why might that happen is not really the point is it? If I ask for a photorealistic image of a man sitting at a computer, a priori I might think 'in what world would I expect seven fingers and no thumbs per hand', alas...
Re:New LLM jailbreak bypasses all major FMs
by BriggyDwiggs42 on Saturday April 26, 2025 @12:59AM
(#43801071)
I’ll take the example as an example of an LLM initiating harmful behavior in general and admit that such a thing is perfectly possible. I think the issue is down to the degree to which preventing such initiation impinges on the agency of the user, and I don’t think that requests for information should be refused because it’s lots of imposition for very little gain. I’m perfectly alright with conditioning/prompting the model not to readily jump into serious, potentially harmful targets without the direct request of the user.