open-compass
[ICLR 2025] xFinder: Large Language Models as Automated Evaluators for Reliable Evaluation