Image: Prompt engineering workshop banner. Generation: DALLE
This is a repository containing strategies I have been experimenting with to improve the outputs from GPTs (but especially ChatGPT) using various attempts to optimise prompts (a field which is now commonly referred to as "prompt engineering").
I will also endeavor to use this repo to gather third-party prompt engineering resources and best practices.
Experiments
- Logs of experiments designed to test the effectiveness of different prompting strategiesGPT_Outputs
- Asking GPT to provide its own prompt eng guidesPrompt Snippets
- Snippets I've found helpful in refining promptsPrompts
- Longer promptsResources
- Mentioned below. Other people's prompt engineering guides.Templates
- Templates for logging prompt engineering work. Mostly GPT generated to start with my subsequent refinements.Workarounds
- Strategies for achieving better outputs
Prompt engineering resources are listed under resources.
Currently this section has the following entities:
- Articles
- Courses & Certifications
- Github repositories
- Guides
- Newsletters
- Organisations
- Websites
For a more complete set of resources, see the various Awesome repos, including Awesome Prompt Engineering.
Prompt engineering is a young and rapidly-evolving discipline and given the rapidity at which GPTs evolve and change also something of a constantly moving target.
Daniel Rosehill public at danielrosehill dot com