LLMs remain vulnerable to "jailbreaking" through adversarial prompts(link.springer.com)
link.springer.com
LLMs remain vulnerable to "jailbreaking" through adversarial prompts
https://link.springer.com/article/10.1007/s11098-025-02347-3
6 comments
Uhhhh, no... idk who that is but this comment from you is both factually incorrect and slightly insulting. whatever hes saying is the shadow of this.
I don't know what you're reading, but I gave a link to a post on Mastodon, and then I quote directly from it.
Here's the link:
https://sigmoid.social/@raphaelmilliere/114659355740586289
Here is the text from that post:
Despite extensive safety training, LLMs remain vulnerable to “jailbreaking” through adversarial prompts. Why does this vulnerability persist? In a new open access paper published in Philosophical Studies, I argue this is because current alignment methods are fundamentally shallow.
That Mastodon post then links to a paper by the same person, so one assumes they are giving an accurate summary of their own work.
So I don't know what you are claiming is factually incorrect.
Here's the link:
https://sigmoid.social/@raphaelmilliere/114659355740586289
Here is the text from that post:
Despite extensive safety training, LLMs remain vulnerable to “jailbreaking” through adversarial prompts. Why does this vulnerability persist? In a new open access paper published in Philosophical Studies, I argue this is because current alignment methods are fundamentally shallow.
That Mastodon post then links to a paper by the same person, so one assumes they are giving an accurate summary of their own work.
So I don't know what you are claiming is factually incorrect.
https://sigmoid.social/@raphaelmilliere/114659355740586289
"Despite extensive safety training, LLMs remain vulnerable to “jailbreaking” through adversarial prompts. Why does this vulnerability persist? In a new open access paper published in Philosophical Studies, I argue this is because current alignment methods are fundamentally shallow."