Enabling Language Models to Implicitly Learn Self-Improvement

Ziqi Wang, Le Hou, Tianjian Lu +4

Large Language Models (LLMs) have demonstrated remarkable capabilities in open-ended text generation tasks. However, the inherent open-ended nature of these tasks implies that ther…