taniakoh/

code-switch-xnli-test

	CS-XNLI: Synthetic Code-Switched NLI Evaluation Corpus




	
	Dataset Summary

CS-XNLI is a synthetically generated code-switched dataset derived from the standard Cross-Lingual NLI (XNLI) evaluation benchmark. This dataset aims to address the scarcity of mixed-language resources for complex reasoning tasks.

1000 Annotated Sentence Pairs for Intra-Sential Code-Switching Each: English-Spanish(es), English-Vietnamese(vi), English-Mandarin(zh)