Unveiling the Misuse Potential of Base Large Language Models via In-Context Learning

Open in new window