What is a Prompt?
A prompt is the input you provide to an AI model to generate a response. In Arato, prompts are at the core of experimenting with different models, parameters, and datasets. A well-structured prompt improves the accuracy, relevance, and consistency of AI-generated outputs.
Example Prompt:
This prompt instructs the model to generate a summary of the provided text.
What is a Variable?
Variables act as placeholders that can be replaced dynamically with different values from a dataset, enabling batch testing across multiple inputs.
Why Using Variables in your Prompt?
To make prompt testing scalable and efficient, use variables within prompt to make it a Prompt Template. There are few key benefits for using variables in prompts:
Automated Testing: Run multiple test cases without manually changing prompts.
Scalability: Test LLM responses on hundreds of variations efficiently.
Comparative Analysis: Evaluate performance across different datasets.
Reproducibility: Maintain consistency in experiments.
How to use Variables in Prompts
Use {{variable}} syntax for variables in prompts.
Example Prompt Template with Variables:
Here, {{text}} is a variable that will be replaced with actual input data from the Dataset.