arXiv · 2412.16683
Dynamical Behaviors of the Gradient Flows for In-Context Learning
Abstract
We derive the system of differential equations for the gradient flow characterizing the training process of linear in-context learning in full generality. Next, we explore the geometric structure of the gradient flows in two instances, including identifying its invariants, optimum, and saddle points. This understanding allows us to quantify the behavior of the two gradient flows under the full generality of parameters and data.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Songtao Lu, Yingdong Lu, Tomasz Nowicki. 2024-12-21. Dynamical Behaviors of the Gradient Flows for In-Context Learning. https://arxiv.org/abs/2412.16683
Cite the original work for its findings. Save a collection to share your selection of sources.