StablePT: Towards Stable Prompting for Few-shot Learning via Input Separation

Xiaoming Liu, Chen Liu, Zhaohan Zhang +4

Large language models have shown their ability to become effective few-shot learners with prompting, revolutionizing the paradigm of learning with data scarcity. However, this appr…