Submitted by Kuei-Chun Kao 20 AutoRubric-T2I: Robust Rule-Based Reward Model for Text-to-Image Alignment Arena 8 1