Researchers Uncovers New Methods To Defend AI Models Against Universal Jailbreaks
Researchers from the Anthropic Safeguards Research Team have developed a new approach to protect AI models from universal jailbreaks. This innovative method, known as Constitutional…