Raemon comments on Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations

[Raemon]

This reminds me of my pet crusade that Abstracts should be either Actually Short™, or broken into paragraphs