Sometimes you'll want to annotate individual examples, but see them in the context of other examples.
This is most common in chatbot annotation, where you annotate individual messages one by one, but want to show the entire conversation to the annotator for context, and display it in order.
Another common use case is annotating long texts for machine learning. It's common practice to split the documents into sentences or paragraphs, but still very helpful for annotators to see the entire article and annotate it in one shot.
For both of these cases, LightTag has contextual display. In contextual display mode, we'll display multiple examples together and in the order you specify.
Setting it up
Contextual display is configured on Datasets at upload time.
In the organize your dataset section
- Toggle the group and order section
- Select an Aggregation Field (this tells lighttag what to group your examples by)
- Select an Order Field (this tells LightTag how to sort examples within a group )
We'll show you a preview on the left
Annotating With Contextual Display
Your annotators will be shown a list of examples in order, one list for each group. When they've annotated each example the submit button at the top will activate and they can move on to the next task