New AI Jailbreak Method 'Bad Likert Judge' Boosts Attack Success Rates by Over 60%

New AI Jailbreak Method 'Bad Likert Judge' Boosts Attack Success Rates by Over 60%
By: The Hacker News Posted On: January 03, 2025 View: 5

Cybersecurity researchers have shed light on a new jailbreak technique that could be used to get past a large language model's (LLM) safety guardrails and produce potentially harmful or malicious responses. The multi-turn (aka many-shot) attack strategy has been codenamed Bad Likert Judge by Palo Alto Networks Unit 42 researchers Yongzhe Huang, Yang Ji, Wenjun Hu, Jay Chen, Akshata Rao, and

Read this on The Hacker News BitAddax - Win with Crypto
  Contact Us
  Follow Us
  About

Headlines Portal is your source for all news globally.