Skip to yearly menu bar Skip to main content


Poster
in
Workshop: 2nd Workshop on Generative AI and Law (GenLaw ’24)

Rethinking LLM Memorization through the Lens of\\Adversarial Compression

Avi Schwarzschild · Zhili Feng · Pratyush Maini · Ethan Herenstein · Zachary Lipton


Abstract:

We propose a new definition of memorization for LLMs based on our Adversarial Compression Ratio with the express goal of offering a metric to be used as evidence in copyright and fair use cases. We argue that if a model can output a piece of text exactly with a prompt that has less information in it, then the model has memorized that data.

Chat is not available.