What Happened in LLMs Layers when Trained for Fast vs. Slow Thinking: A Gradient Perspective Paper • 2410.23743 • Published Oct 31, 2024 • 59
