Multilingual

HELM CLEVA (Chinese)

Chinese language evaluation across 21 released task configurations and four model exports. Preserve the declared task metric: item grades where published, and null grades for attempts evaluated only by corpus-level metrics.