Paper: Accepted in LREC 2026Contact:
ekhi.azurmendi@ehu.eusThe HABE-HiTZ C1 dataset is a collection of essays and annotations designed for training and evaluating models on scoring and feedback generation.
The dataset is organized into three distinct subsets. To ensure traceability, every essay is assigned a unique ID that remains consistent across all three subsets, allowing you to easily map scores, feedbacks, generations and… See the full description on the dataset page:
https://huggingface.co/datasets/EkhiAzur/HABE-HiTZ_C1_Dataset.
