2025

• R1 Reasoning - March 2025

2024

• **Reward Hacking, Shortcut Learning, and Spurious Correlation** - December 2024