What Are Large Language Models Capable Of: The Vulnerability of LLMs to Adversarial Attacks

Recent research uncovered a vulnerability in deep learning models, including large language models, called "adversarial attacks." These attacks manipulate input data to mislead models. So, I decided to test out a framework that automatically generates universal adversarial prompts.

Read All

adversarial adversarial attacks ai ai ethics ai vulnerabilities attacks data deep learning ethical ai framework future-of-ai language language models large language large language models llms prompts research test vulnerability