AI Red Teams for Adversarial Training: Making ChatGPT and LLMs More Robust(surgehq.ai)

https://www.surgehq.ai//blog/ai-red-teams-for-adversarial-training-making-chatgpt-and-large-language-models-adversarially-robust