Transformers Learn Higher-Order Optimization Methods for In-Context Learning: A Study with Linear Models

Open in new window