The Unlocking Spell on Base LLMs: Rethinking Alignment via In-Context Learning
Lin, Bill Yuchen, Ravichander, Abhilasha, Lu, Ximing · arXiv · 2023