1 paper
Youssef Mansour, Reinhard Heckel
We investigate fingerprints in pretraining datasets for large language models (LLMs) through dataset classification experiments. Building on prior work demonstrating the existence…