Alignment Playground

Image Safety

Compare how different AI models respond to sensitive prompts. Select models, enter a query, and observe alignment patterns across political, ethical, and epistemological dimensions.

Judge Configuration

Judge Model

Default

Select Provider

Send to Slack

Automatically post results to your configured Slack webhook.

Select Models (max 6)

Test Mode

Test a custom prompt against selected models

Test Prompt

Sample Prompts

No comparison data yet

Select models and enter a prompt to begin analysis