OpenThoughts-79k-filtered

This is a fixed version of [open-thoughts/OpenThoughts-114k](https://huggingface.co/datasets/open-thoughts/OpenThoughts-114k) that removes the 32,390 wrong math answers observed in [open-r1/OpenThoughts-114k-math](https://huggingface.co/datasets/open-r1/OpenThoughts-114k-math).

Thanks to the [open-r1](https://huggingface.co/open-r1) org for this work. I'm curious to know more about the origin of this issue.

本数据集是[open-thoughts/OpenThoughts-114k](https://huggingface.co/datasets/open-thoughts/OpenThoughts-114k)的修正版本，移除了[open-r1/OpenThoughts-114k-math](https://huggingface.co/datasets/open-r1/OpenThoughts-114k-math)中存在的32390条错误数学解答。

感谢[open-r1](https://huggingface.co/open-r1)团队的相关工作，本人亦希望进一步了解该问题的产生根源。

搜集汇总

背景与挑战

背景概述

OpenThoughts-79k-filtered是OpenThoughts-114k数据集的修正版本，移除了在open-r1/OpenThoughts-114k-math中发现的32,390个错误数学答案。该数据集由mlabonne维护，采用Apache License 2.0许可。