Adding Dataset
Add from Existing Dataset or Experiment
This feature allows users to incorporate data from previously created datasets or experiments, streamlining the workflow by reusing structured information. Instead of manually recreating datasets, you can efficiently map and integrate existing data, ensuring consistency and reducing redundancy.
Steps to Add from an Existing Dataset or Experiment
1. Access the Dataset Selection Panel
- Navigate to the Datasets & Experiments section.
- Click “Add Dataset” to open the dataset creation menu.
- Select “Add from Existing Model Dataset or Experiment” from the available options.
2. Select the Source Dataset
- A panel will display a list of available datasets and experiments.
- Use the search bar to quickly locate a dataset by name.
- Choose the dataset containing the relevant data.
3. Configure Column Mapping
- After selecting the dataset, a “Map to New Column” interface appears.
- Align columns from the existing dataset with corresponding fields in the new dataset.
- Ensure proper mapping of Input, Context, Output, Expected Response, and Metadata for accuracy.
4. Import
- Choose between:
- “Import Data” – Directly imports the dataset’s values.
- “Import Data and Prompt Configuration” – Imports data along with existing prompt settings.
- Click “Add” to confirm and integrate the dataset.
Key Benefits
- Efficiency – Saves time by utilising pre-existing datasets.
- Consistency – Ensures uniform data formatting and structure.
- Scalability – Supports seamless dataset expansion.