LƦ | Ʀamblings

Large Language Models Prompting For Academic Research Code

Academic research often produces low-quality code due to its primary purpose of hypothesis testing and data analysis. Once published, this code usually becomes abandoned.

Today, thanks to LLMs, we can write high-quality, robust functions through smart prompting techniques.

For example, providing a structured input and carefully designed system prompts, a local LLM can generate testable Python functions.

This approach will improve the way researchers develop and utilize code, preventing the dreadful phenomenon of abandonware that plagues academia.

Moreover, this approach constrains the thought process into computational-thinking. Therefore, researchers can refine their thought process and , at the same time, improve the way they design functional code!