🏷️Tags
Last updated
Last updated
Tags serve as a fundamental feature within our platform. Each tag functions as an identifier, enabling users to group prompt runs and responses for easier access and analysis. You can choose either a preexisting tag from the dropdown list on the Prompt Workshop page, or begin typing to create a new tag. If you choose not to add tags initially, you will be able to do so later in the Prompt Runs page after running your prompt. Additionally, when you choose the scorer(s) you want to run your prompt against, its scorer tag is automatically attached to the prompt run. Below in this document, there are screenshots of where you can view and modify tags.
Identification and Grouping: Each tag functions as an identifier, enabling users to group prompt runs and responses for easier access and analysis.
Multiple Tags per Response: A single response can have multiple tags, offering a versatile way to classify data across different dimensions like experiments, models, results, etc.
Customization Options: Tags are customizable in two ways.
Unique Name: Every tag should have a unique name to ensure clear identification.
Color Assignment: Users can assign a color to each tag, enhancing visual categorization and recognition.
Tags can be assigned within various pages on the UI:
Prompt Workshop: When creating a new prompt, users can assign tags directly on the Prompt Workshop page. This practice is crucial for later identifying and distinguishing between prompts within an experiment. We recommend adding tags that represent template and experiment levels, as well as model levels, to every prompt run. In the screenshot below, tags are outlined by the red box. The tags "CEOs" and "Example" were created by the user by simply clicking into the Editor field and typing the tag name on the popup box. Colors can also be assigned promptly after submitting the tag name.
Prompts Runs: Tags can also be assigned to prompt runs in bulk on the Prompts Run page. This functionality is particularly useful for organizing data to be analyzed in the monitoring and analytics sections. In the screenshot below, tags are outlined by the red box. The tags "CEOs" and "Example" were created by the user. The "GPT-4" and "MISTRAL-TINY" tags were automatically attached by Evaluable AI since those prompts were run against those respective models. To add more tags, simple click on the "+" and type the tag name on the popup. Colors can also be assigned promptly after submitting the tag name.
We are exploring the addition of automatic tags for every prompt run to automatically identify and group tags based on templates and models aiming to further simplify the process of identifying and classifying prompts without the need for manual tagging by users.
