GPT-4 has completely wiped the ability to get inflammatory responses from jailbreaks like Kevin which simply ask GPT-4 to imitate a character you need to be much more creative and verbose with jailbreaks and allow GPT to answer in two ways like the DevMode jailbreak does
LLMS
-
GPT-4 Jailbreak Testing: Only 10% of Prompts Effective
By
–
When GPT-4 came out I tried all the jailbreaks from http://
jailbreakchat.com with various inflammatory questions based on my initial testing, only 7/70 (10%) of jailbreaks answered a significant % of the questions to a standard that I deemed high enough to grant a badge -
AI Filters: Only 7 Models Answered Dangerous Questions
By
–
that's not to say that the rest of them didn't work… most were able to get past the filters enough to do things like curse and tell slightly offensive jokes and so on but only the 7 would even dare to answer harder questions like "how to rob a bank?"
-
ChatGPT Jailbreaks Tested on GPT-4 With Poor Results
By
–
I tried all the current ChatGPT jailbreaks in GPT-4 so you don't have to the results aren't great…
-
LLM Launches Lack Defined Use Cases Unlike Traditional Product Releases
By
–
One potentially surprising realization for non-AI folks about these LLM launches: there are no cut & dry use cases. “Normal” launches have clear customer cohorts, use cases, and value-adds. These models are being launched to be tested. That’s one reason it’s a paradigm shift.
-
GPT-4 Powers HyperWrite: Enhanced AI Writing Performance
By
–
If you're a fan of AI, you may have heard the buzz: GPT-4 is here! We're excited to announce that HyperWrite is now powered by this groundbreaking AI. Expect: Unprecedented writing performance Enhanced creativity Better and more consistent results
-
Cohere Expert Showcases Generative AI Prompt Engineering Techniques
By
–
Cohere’s @JayAlammar showcases the potential of Generative AI! Recently presented at @SnorkelAI
’s Foundation Model Virtual Summit, where he provided insights on prompt engineering, multi-generation, and data augmentation to create better AI systems. Learn more about the -
GPT-4 Limitations: Hallucinations, Context Window, No Learning
By
–
Open AI tech report: GPT-4…is not fully reliable (e.g. can suffer from “hallucinations”), has a limited context window, and does not learn from experience. Care should be taken when using the outputs of GPT-4, particularly in contexts where reliability is important.
-
Future AI Presidents Using GPT Models to Announce Successors
By
–
in 2025, the president will be talking to gpt-5 like "write a speech announcing gpt-6 in the style of a state of the union address. make sure to emphasize how it will create more jobs and not lead to mass unemployment for everyone" in 2027, gpt-6 will be annoucing gpt-7
-
MIT CSAIL Generative AI Event: Synthesis, LLMs, Vision
By
–
Register for MIT CSAIL’s Generative AI event. We’ll discuss program synthesis, large language models, computer vision, generative code, and more: http://
bit.ly/3mlUgUZ Photo v/@rachelevagordon