lxyuan/

nemo-codeswitch-reasoning-debate

This is a synthetic, multilingual code-switching dataset. Each record contains:

a realistic user query a long-form reasoning section a debate / counterargument section a concise final_answer

It is designed for experiments in multilingual generation, code-switch robustness, and reasoning/debate style responses. This snapshot contains 574,977 rows and 10 string columns.

Important:

Verify that your intended usage and redistribution complies… See the full description on the dataset page: https://huggingface.co/datasets/lxyuan/nemo-codeswitch-reasoning-debate.