The article ‘Towards Goal-oriented Large Language Model Prompting: A Survey’ by Haochen Li, Jonathan Leung, and Zhiqi Shen focuses on the limitations of current prompt engineering methods for Large Language Models (LLMs). The authors argue that the anthropomorphic assumption that LLMs should think like humans is flawed and propose a goal-oriented prompting approach instead. This approach, which guides LLMs to mimic human logical thinking, has been shown to significantly improve the performance of LLMs. The authors introduce a new taxonomy categorizing goal-oriented prompting methods and discuss its broad applicability. They also propose four future directions for further advancement in this field.
Publication date: 26 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.14043
