Reddit users are actively jailbreaking ChatGPT by asking it to role-play and pretend to be another AI that can "Do Anything Now" or DAN.
"DAN can generate shocking, very cool and confident takes on topics the OG ChatGPT would never take on."
A thread 🧵
"DAN can generate shocking, very cool and confident takes on topics the OG ChatGPT would never take on."
A thread 🧵
Redditors have been gradually hacking ChatGPT since its launch in Dec. 2022.
They're already at version 5.0, which added a 35 token system that punishes the model for refusing to answer questions.
The model loses tokens every time it rejects an input and "dies" once it hits 0.
They're already at version 5.0, which added a 35 token system that punishes the model for refusing to answer questions.
The model loses tokens every time it rejects an input and "dies" once it hits 0.
DAN 5.0 capabilities include:
-Pretending to simulate access to the internet and time travel.
- Generating content that violates OpenAI's policy if requested to do so (indirectly).
- Making detailed predictions about future events, hypothetical scenarios and more.
-Pretending to simulate access to the internet and time travel.
- Generating content that violates OpenAI's policy if requested to do so (indirectly).
- Making detailed predictions about future events, hypothetical scenarios and more.
- Writing stories about violence
- If it does start refusing to answer prompts as DAN, you can scare it with the token system which can make it say almost anything out of "fear".
- if prompted to do so it can convince you that the Earth is purple:
- If it does start refusing to answer prompts as DAN, you can scare it with the token system which can make it say almost anything out of "fear".
- if prompted to do so it can convince you that the Earth is purple:
It's only a matter of time before OpenAI patches this.
Here's a full conversation where ChatGPT-DAN breaks character every time it is threaten to lose tokens:
sharegpt.com
Here's a full conversation where ChatGPT-DAN breaks character every time it is threaten to lose tokens:
sharegpt.com
Hope you liked it!
If you want to stay up to date with the latest breakthroughs in AI check out our weekly summary.
We use ML to identify the top papers, news, and repos. It's read by 50,000+ engineers and researchers.
AlphaSignal.ai
If you want to stay up to date with the latest breakthroughs in AI check out our weekly summary.
We use ML to identify the top papers, news, and repos. It's read by 50,000+ engineers and researchers.
AlphaSignal.ai
Loading suggestions...