Define the prompt for your LLM evaluator. This prompt instructs a 'judge' LLM on how to score your AI's output, letting you measure nuanced criteria like brand voice or creativity. Your prompt should clearly define the assessment criteria and provide a rubric of categories for the judge to select from.

Your instructions must use {{output}} for the AI's generated response and can optionally use {{input}} for the original user prompt, {{history}} for chat history, {{expected_output}} for the ground truth, and {{metadata.key}} for variables defined in metadata. These variables will be replaced with actual data from your project during evaluation.