Exploring different techniques for using LLMs to evaluate jailbreak attempts
Finetuning of Mistral Nemo 13B on the WildJailbreak dataset to produce a red-teaming model