Critic: Large Language Models Can Self-Correct with Tool-Interactive Critiquing

4 points| fyskij | 2 years ago |arxiv.org