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.
Successful Page Load