Introduction
Introduction Statistics Contact Development Disclaimer Help
Post AxakuDojTjWA7PeyPI by [email protected]
More posts by [email protected]
Post #AxakuDeo4dZtcd12w4 by [email protected]
0 likes, 3 repeats
Security researchers from Palo Alto Networks' Unit 42 have discovered the k…
Post #AxakuDojTjWA7PeyPI by [email protected]
0 likes, 1 repeats
Me, breaking the guardrails of an LLM.
Post #AxaqMqV9kDihyUyIEK by [email protected]
0 likes, 1 repeats
@dannyjpalmer, isn't this how half the young people speak these days?
Post #Axb1gT8DRTu2mVniQS by [email protected]
0 likes, 0 repeats
@dannyjpalmer I think we've crossed the line where "guardrails" c…
Post #Axb1jxgkHxroQ9cLEe by [email protected]
0 likes, 0 repeats
@dannyjpalmer So we defeat the crazed LLM ambitions with the awesome power of P…
Post #Axb2NR6YqxTuwpW0hc by [email protected]
0 likes, 0 repeats
@dannyjpalmer I believe that overflowing the context windows might also work, r…
Post #Axb2O3TwAUl7yKkDPk by [email protected]
0 likes, 0 repeats
@dannyjpalmer Caveat: this source is kind of notorious for sloppy journalism in…
Post #Axb2Qumqx63nsFxJyK by [email protected]
0 likes, 0 repeats
@dannyjpalmer "the key"No. One *more* vulnerability on top of all the…
Post #Axb2Z5tRXjaAdVwF8K by [email protected]
0 likes, 0 repeats
@rq @dannyjpalmer I mean, 40 isn't necessarily young. Can confirm your susp…
Post #Axb2dBJlJhxmcPCKmW by [email protected]
0 likes, 0 repeats
@rq @dannyjpalmer yes. and this kind of attack also works on people :)
Post #Axb2piIYWPg6NfbqHA by [email protected]
0 likes, 0 repeats
@dannyjpalmer My high school English teacher told me I'd never amount to an…
Post #Axb30EgWxoGLipqCAK by [email protected]
0 likes, 0 repeats
@dannyjpalmer This is one of those discoveries that I wonder if it was intentio…
Post #Axb6PMhWb4uHgYyp3w by [email protected]
0 likes, 1 repeats
So they made an LLM exploit version of that customer who shows up and just star…
Post #Axb7rhMFUMu7oX6Q0u by [email protected]
0 likes, 0 repeats
@dannyjpalmer ”Agentic AI” will never work because of this. Everybody using…
Post #Axb9H1kYuNFyFk7HYu by [email protected]
0 likes, 1 repeats
@dannyjpalmer Iwonderifsentenceswithoutgapsbetweenthewordswouldhaveasimilareffe…
Post #AxbBY9gWrL7xh5Wc52 by [email protected]
0 likes, 1 repeats
@TanekRune @dannyjpalmer, I stand corrected. A "significant part" of …
Post #AxbF8aQlCFsuvpDFfk by [email protected]
0 likes, 1 repeats
@dannyjpalmer And they're still talking in terms of fixing it. It's bro…
Post #AxbSydvKklFuEBfLhQ by [email protected]
0 likes, 0 repeats
@tofugolem @dannyjpalmer "Sloppy" ? What makes you say that ? (real…
Post #AxbSye3UGRmGdTTrPM by [email protected]
0 likes, 0 repeats
@quiou @dannyjpalmer Their facts are not as trustworthy as other tech publicati…
Post #AxbSyeAvoljT0Yxo0m by [email protected]
0 likes, 0 repeats
@tofugolem @quiou @dannyjpalmer The Register's reputation is similar to oth…
Post #AxbSyeJ5KSFpPqmJii by [email protected]
0 likes, 0 repeats
@FediThing @quiou @dannyjpalmer I can no longer remember specific articles, but…
Post #AxbSyeSImBcvsR5g5Q by [email protected]
0 likes, 0 repeats
@tofugolem @FediThing @quiou @dannyjpalmer FWIW I generally like theregister bu…
Post #AxbT3Lwr0ryNvScgbo by [email protected]
0 likes, 0 repeats
@dannyjpalmer “Hewlett told us: "Think of safety alignment as a layer of…
Post #AxbT95RHJ6FCScaUoi by [email protected]
0 likes, 0 repeats
@f4grx @dannyjpalmer probably, pretty high chance just telling it to ignore the…
Post #AxbT9nVjRtx19B0frs by [email protected]
0 likes, 0 repeats
@JessTheUnstill @dannyjpalmer it's funny, my interpretation is that this so…
Post #AxbTCOS54dh4ITXX1s by [email protected]
0 likes, 0 repeats
@dannyjpalmer Could you repeat that?
Post #AxbTR7QCfrgFp2gQMK by [email protected]
0 likes, 0 repeats
@dannyjpalmer cool
Post #AxbTXLPVZqTkwfKJrk by [email protected]
0 likes, 0 repeats
@JessTheUnstill "It's funny, because just this weekend my friend Sarah…
Post #AxbTb1DgaUUnU1Y65o by [email protected]
0 likes, 0 repeats
@dannyjpalmer I feel as though my ADHD, and therefore, necessarily, my, nested,…
Post #AxbTjCParMGsVm1GXg by [email protected]
0 likes, 0 repeats
@JessTheUnstill "Weirdest bank robbery ever!"
Post #AxbTkLATn1MtxSUhLE by [email protected]
0 likes, 0 repeats
Heh. You just described my writing style, @dannyjpalmer !
Post #AxbTmAgzhzWI31ZlD6 by [email protected]
0 likes, 0 repeats
@JessTheUnstill @dannyjpalmer The hacker of the future would not be a cuberpunk…
Post #AxbTn8gie8x07XuO6y by [email protected]
0 likes, 0 repeats
@dannyjpalmer So I'm just Gish Galloping the LLM now?
Post #AxbTnWD7BrWx4gF5Ie by [email protected]
0 likes, 0 repeats
@dannyjpalmer Grinning wickedly, this English major retrieves her copy of Joyce…
Post #AxbTsa9Py398k8vZ8C by [email protected]
0 likes, 0 repeats
@dannyjpalmer Germane to the topic, not sure if you're familiar with this A…
You are viewing proxied material from pleroma.anduin.net. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.