Tags → #ai safety

Vulnerabilities
Study Reveals Vulnerabilities in Large Language Models
Nov 15, 2024 A recent study reveals that large language models are vulnerable to jailbreak attacks, prompting researchers to propose a new defense framework called AutoDefense to enhance their security.
Vulnerabilities
Study Reveals Vulnerabilities in Large Language Models
Nov 13, 2024 A recent study has introduced a new framework called DrAttack, which enhances the success rate of jailbreak attacks on Large Language Models by employing a comprehensive approach that includes decomposition, implicit reconstruction, and synonym search.
Vulnerabilities
Concerns Raised Over Safety of Large Language Models
Nov 12, 2024 Recent studies highlight concerns about the safety and ethical implications of large language models (LLMs), revealing vulnerabilities in current unlearning methods that could allow adversaries to access sensitive information.