Abstractlarge language models (LLMs) have exhibited impressive capabilities in various domains, particularly in general language understanding. However these models, trained on massive text data, may not be finely optimized for specific tasks triggered by instructions.
→