Picture for Raphaël Sarfati

Raphaël Sarfati

Density estimation with LLMs: a geometric investigation of in-context learning trajectories

Add code
Oct 07, 2024
Viaarxiv icon

Lines of Thought in Large Language Models

Add code
Oct 02, 2024
Viaarxiv icon

LLMs learn governing principles of dynamical systems, revealing an in-context neural scaling law

Add code
Feb 01, 2024
Viaarxiv icon