Dataset Access Notice
To obtain access to the dataset, interested users must first complete a registration on the Codebench platform. Following successful registration and approval, applicants will receive access to the training datasets for both English and Spanish.
The dataset used in this workshop is structured into three main classification tasks, each designed to capture a different dimension of social support in online discourse. The distribution of samples across tasks and categories is summarized below. These numbers reflect the total annotated instances used for training and evaluation of the classification models.
Task 1 – Social Support Detection
This task distinguishes whether a comment expresses social support or not.
Social Support: 2,236 samples
Not Social Support: 7,762 samples
Task 2 – Type of Support Recipient
This task identifies whether the support is directed toward an individual or a group.
Individual: 423 samples
Group: 1,813 samples
Task 3 – Targeted Communities and Identities
This task classifies the target of support into specific community or identity-based categories.
Nation: 982 samples
Other: 520 samples
LGBTQ: 154 samples
Black Community: 114 samples
Women: 24 samples
Religion: 19 samples