Our paper Inference-Time Reward Hacking in Large Language Models was selected as a spotlight at the NeurIPS 2025.