https://github.com/CivicActions/prompts

CivicActions Generative AI bot prompts and configuration
https://github.com/CivicActions/prompts

Last synced: 21 days ago
JSON representation

CivicActions Generative AI bot prompts and configuration

• Host: GitHub
• URL: https://github.com/CivicActions/prompts
• Owner: CivicActions
• License: gpl-3.0
• Created: 2024-02-16T01:55:40.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2024-02-16T03:01:35.000Z (about 1 year ago)
• Last Synced: 2025-01-28T03:31:50.966Z (25 days ago)
• Size: 19.5 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
This is used to manage bot and prompt configuration for CivicActions Slack chatbot.

All configuration is in the [config.yaml](config.yaml) file. There is a default configuration, then a configuration for additional bots that are available.

Each bot can include the following parameters, all optional:

* `description`: A message describing the bot behaviour. If this is a prompt created elsewhere include a link and license.

* `prompt`: Override a default or earlier prompt.

* `prompt_extra`: Extend a default or earlier prompt by adding this on the end. Generally preferred over `prompt` as this provides for additional flexibility in composing prompts.

* `temperature`: Set the temperature between 0 and 2 - the controls how much randomness is incorporated into the generative algorithm. Higher values tend to lead to more "creative" output.

* `max_tokens`: Cap the number of tokens to be output. Users can ask the bot to "continue" if more output is desired.

Users are able to compose multiple bots in order - e.g. `bot1+bot2` - each parameter overrides the default or earlier parameters, except for prompt_extra which will append to the previous prompt.

The following parameters are also available, but generally are only used in specific situations.

* `model`: The provider to use (needs to be preconfigured)

* `deployment`: The specific deployment at the provider to use (needs to be preconfigured)