Sophos AI to present on how to defang malicious AI models at Black Hat Europe

“LLMbotomy” research reveals how Trojans can be injected into Large Language Models, and how to disarm them.