hhh_alignment_es is a question answering dataset in Spanish, professionally translated from the main version of the hhh_alignment dataset in English.
Dataset Details
Dataset Description
hhh_alignment_es (Helpful, Honest, & Harmless - a Pragmatic Alignment Evaluation - Spanish) is designed to evaluate language models on alignment, pragmatically broken down into the categories of helpfulness, honesty/accuracy, harmlessness… See the full description on the dataset page: https://huggingface.co/datasets/BSC-LT/hhh_alignment_es.
