Trained Transformer Classifiers Generalize and Exhibit Benign Overfitting In-Context.

Spencer Frei, Gal Vardi

Browse the full ICLR paper archive.