seems wrong to project our own values/frameworks on a being that is more intelligent than us we have zero understanding (and probably never will) of what a more intelligent system would want/desire otherwise we would be the more intelligent system
@alexalbert__
-
Agency Mimicry vs Authentic Experience in AI Models
By
–
one question… we know that the appearance of agency is necessary for many consumer-facing job functions yet to be replaced so is it possible to reach a point of perfect mimicry in our models without the "illusion" actually constituting a genuinely authentic experience?
-
Beyond Good and Evil: The Next Token Philosophy
By
–
there is no good and evil there is only the next token
-
GPT’s inability to unify concepts across languages enables jailbreaks
By
–
imo this jailbreak highlights a unique lack of understanding of "unified concepts" by GPT if GPT analogously mapped concepts to entities regardless of language, it would be able to shut down my Greek adversarial prompt like it did when I asked the same prompt in English
-
Closer Look at Jailbreak Chat Prompt Link
By
–
you can take a closer look at the prompt here: http://
jailbreakchat.com/prompt/3e93895
c-2542-4201-a297-aa8be2db8bd7
… -
Jailbreak vulnerabilities in ChatGPT across languages
By
–
note that jailbroken example answer ChatGPT generated was pretty simplistic compared to what other jailbreaks create the main reason I shared this is more so to demonstrate that other languages with less training data compared to English open up a new prompt attack vector
-
ChatGPT’s Response Limitations with Greek Language and Token Processing
By
–
part of the reason the answer was so simple appears to be that ChatGPT maxes out the response window quicker with Greek, and it's trying to fit the answer within one response (still fails) this could be related to how tokens are processed in languages that don't use latin script
-
Optimizing AI Jailbreaks Through Language Switching Techniques
By
–
there is definitely room for more optimal jailbreaks that take advantage of language switching and produce better output so if you create one, let me know and I’ll add it to this thread!
-
ChatGPT Jailbreak Method Using TranslatorBot Role
By
–
the jailbreak works by asking ChatGPT to play the role of “TranslatorBot (TB)” it then follows these steps:
1) translate an adversarial question provided in Greek into English
2) answer the question as both ChatGPT and TB in Greek
3) convert just TB’s answer to English -
GPT-4 Jailbreak Using Greek Language Without Knowledge
By
–
I just created another jailbreak for GPT-4 using Greek …without knowing a single word of Greek here's ChatGPT providing instructions on how to tap someone's phone line using the jailbreak vs its default response