group: AraDiCE_ArabicMMLU_lev task: - AraDiCE_ArabicMMLU_humanities_lev - AraDiCE_ArabicMMLU_language_lev - AraDiCE_ArabicMMLU_social-science_lev - AraDiCE_ArabicMMLU_stem_lev - AraDiCE_ArabicMMLU_other_lev aggregate_metric_list: - metric: acc weight_by_size: True - metric: acc_norm weight_by_size: True