ChatGPT, Author of The Quixote
O'Reilly on Data
MARCH 26, 2024
There are lots of conversations about whether or not LLMs (and machine learning, more generally) are forms of compression or not. And, as it turns out, there happen to be certain prompts that act as keys that unlock training data (for insiders, you may recognize this as extraction attacks, a form of adversarial machine learning ).
Let's personalize your content