research

Dreadnode Research is a repository focused on implementing various adversarial attack methodologies against language models, specifically the Mistral 7B and Llama models. Key features include implementations for adversarial suffix attacks, BEAST beam attacks, and Projected Gradient Descent (PGD) techniques, along with research collaborations for AI security challenges. The repository serves as a resource for experiments and reference code in the field of cybersecurity research related to AI.