User Guide
- Manage Prompts
- Create Prompt
- Edit Version
- Generate Version
- Rename Version
- Delete Version
- Set Primary Version
- Compare Versions
- Token Count and Cost
- Update Prompt Use Case
- Copy Prompt to Workspace
- Move Prompt to Workspace
- Optimize Prompts
- Prompt Scoring
- Scoring Criteria
- Prompt Risk
- Prompt Balance
- Prompt Heatmap
- Connect AI Providers
- AI21 Labs
- Amazon Bedrock
- Anthropic
- Azure OpenAI
- Azure Serverless
- Cohere
- Deep Infra
- Fireworks AI
- Google Gemini
- Groq
- Hugging Face
- Hyperbolic
- Lambda
- Meta
- Mistral
- Moonshot AI
- OpenAI
- Perplexity
- Together AI
- xAI
Prompt Heatmap
On this page you'll find the following:
Get Started
Prompt heatmaps are extremely useful as they provide a visual indicator of what the AI paid most attention to in a prompt (or the least). This helps ensure you and the AI model are aligned as to what parts of a prompt are most important.
To get started, go to the version's Heatmap tab and click the purple Generate heatmap button.

This will bring up the Generating Heatmap modal box. After a few moments the modal box will go away and the prompt heatmap will be displayed.

The heatmap will be shown using the selected color palette. You'll also see how many tokens were used and how long it took to generate the heatmap. The right sidebar contains a SUMMARY with information about the attention paid to various phrases, and a legend for how to interpret the colors assigned the various prompt parts.
It's possible not every part of a prompt will have a color. This can be caused by a couple things: 1) The importance of the phrase or sentence was essentially zero. 2) Some smaller models struggle to assign attention weights to all prompt parts due to their limited capacity and reduced parameter count.

If you hover over each of the prompt parts, a tooltip will appear with an explanation for why the phrase or sentence received the level of attention it did.

Heatmap Color Legend
When generating a prompt heatmap, each phrase or sentence is assigned an attention score that corresponds to a color shade. The color shades are as follows:
Color Shade | Description |
---|---|
Very dark | The primary focus and central to the intent. |
Dark | Highly relevant but not the most important. |
Medium | Essential for context but not the focal point. |
Light | Somewhat relevant but less critical for context. |
Very light | Insignificant or negligible contribution. |
Change Color Palette
The default color palette for a prompt heatmap is red-orange. You can change this by selecting one of the other palettes at the top of the page, resulting in heatmaps that look like these:


Generate Heatmap with Different Models
To change which AI model is used to generate the prompt heatmap, click the Settings link near the top-right of the page.

This will open a side panel where you can select different models and include the prompt's examples as part of the analysis. Make your model selection, click the green Apply button, then generate the prompt heatmap.
Including the prompt's examples can greatly increase the time it takes to generate the heatmap. It also does not guarantee the AI model will assign much, if any, attention weight to the examples. This is because some AI models deem examples as tertiary information and not part of the main content.
