On Cartesian Prompt Injection or The Demon on the Machine
中文摘要
本文借用笛卡尔的“恶魔”概念探讨提示词注入，解析AI模型易受欺骗的原因。
English Summary
Exploring prompt injection through Descartes' "evil demon" concept, this article examines why AI models are susceptible to deception.
Original Excerpt
Prompt injection, Descartes, and why models can be deceived Continue reading on Medium »