Peer Review in Large Language Models based on Consistency Optimization

Large language models can evaluate each other's responses in an unsupervised manner, and their final ranking can be optimized to align with human preferences by maximizing the consistency between the models' capabilities and their scores.