Prompts

What is a Prompt?

A prompt is a natural language instruction provided to a generative model to direct its response or produce a desired outcome. It may include questions, commands, contextual details, few-shot examples, or partial inputs for the model to complete or extend.

A prompt typically includes:

• Prompt Template: An instruction containing placeholders.
• Input Arguments: Dynamic inputs that replace the placeholders.

Note: System prompts might not always have Input Arguments.

Additionally, to generate the prompt using the template and input arguments, there should be a:

• Creation Logic: Code that generates, retrieves, or formats the input arguments and fills them into the template.

Note: If a prompt contains no processing or formatting logic, the creation logic can directly return the prompt template.

Managing Prompts on the Platform:

The Prompt Registry organizes all the registered prompts into customized groups at this centralized location and allows easier tracking, monitoring, and creating new ones.

Registering a Prompt:

1. Click on Create button in Prompt Registry.
2. Fill in important details like Name, Attributes (alias), Properties (Description, Group, Permissible Purpose, Approval Workflow).
3. Select registered resources (like Model, Global Functions, etc.) which can help with prompt creation. These resources can be used in the prompt creation logic section.
4. Define the Input Arguments, Prompt Template, Prompt Creation Logic.
5. Add notes, attach documentation if available in the Additional Information section.
6. Lastly, click on the Save button to complete the registration process.

The registered prompts can be evaluated in the Prompt Registry or can be used in downstream objects (like RAG, Model, Pipeline, Reports, etc.).

Benefits of Prompt Registration:

• Automated tracking and recording of modifications with efficient version upgrades.
• Automatic detection of Permissible Purpose violations.
• Perform evaluations using standardized and custom validation kits.
• Enhances reusability across downstream applications and enables usage tracking with Lineage Tracking.
• Journey to production becomes more transparent and fully auditable, and production monitoring gets easier.
• Extract ready-to-productionize executable artifact.