HELM's winogrande_afr scenario evaluates commonsense pronoun resolution using Winogrande items translated into 11 low-resource African languages.
unassessed
| Category | knowledge |
|---|---|
| Subcategory | reasoning |
| Page status | active |
| Metric | exact match |
| Direction | higher_is_better |
| Unit | percent |
| Publisher | Stanford CRFM (HELM integration); translations from the Bridging the Gap project |
HELM's winogrande_afr scenario evaluates commonsense pronoun resolution using Winogrande-style multiple-choice items translated into 11 low-resource African languages, not English.
Multiple-choice pronoun resolution; each item is a sentence with an ambiguous pronoun and two candidate referents, scored by exact match on the correct choice.
No model card in ModelSpec reports this benchmark yet.