GSM8K_distilled_zh is a Chinese dataset designed for mathematical reasoning, which has been processed using MetaMath.
The question-answer pairs within this dataset have been translated from the original GSM8K dataset (available at
https://github.com/openai/grade-school-math/tree/master) utilizing GPT-3.5-Turbo with few-shot prompting techniques.
This dataset comprises 7,473 training samples and 1,319 testing samples. The training samples are intended for supervised… See the full description on the dataset page:
https://huggingface.co/datasets/PaddlePaddle/GSM8K_distilled_zh.
