Two years after ChatGPT hit the scene, there are numerous large language models (LLMs), and nearly all remain ripe for jailbreaks — specific prompts and other workarounds that trick them into ...