Prompting large language models
Prompt engineering refers to tweaking a pre-written prompt - the text context provided to a large language model (LLM) - in order to improve LLM response quality. I’ve used prompt engineering to improve student math Q&A, for example.
This post rounds up a few resources related to prompt engineering.
Resources:
- LLM Prompt Tuning Playbook by Varun Godbole, Ellie Pavlick: https://github.com/varungodbole/prompt-tuning-playbook
- From Ethan Mollick:
- “Innovation through prompting”: https://www.oneusefulthing.org/p/innovation-through-prompting
- “Prompts for Instructors”: https://www.moreusefulthings.com/instructor-prompts
- “Stop Writing All Your AI Prompts from Scratch”: https://hbsp.harvard.edu/inspiring-minds/an-ai-prompting-template-for-teaching-tasks
- There are approaches beyond writing and refining a single prompt that we might use in the future: “You probably don’t know how to do Prompt Engineering” by Allen Roush
- A useful overview and catalog of terms used to refer to various flavors of prompt engineering: https://www.promptingguide.ai/
Example prompts: