Researchers at ETH Zurich created a jailbreak attack that bypasses AI guardrails

Retrieved on: 2023-11-27 21:30:31

Tags for this article:

Artificial intelligence

Large language models

Machine learning

OpenAI

Deep learning

Reinforcement learning from human feedback

ChatGPT

Generative pre-trained transformer

Prompt engineering

GPT-4

Feedback

Generative artificial intelligence

Click the tags to see associated articles and topics

Researchers at ETH Zurich created a jailbreak attack that bypasses AI guardrails. View article details on HISWAI: https://cointelegraph.com/news/researchers-artificial-intelligence-eth-zurich-created-a-jailbreak-attack

Excerpt

... model that relies on human feedback, including the most popular large language models (LLMs), could potentially be jailbroken. Jailbreaking is a ...

Article found on: cointelegraph.com

View Original Article