dolphin-r1-flash

# Dolphin R1 Flash 🐬

An Apache-2.0 dataset curated by [Eric Hartford](https://huggingface.co/ehartford) and [Cognitive Computations](https://huggingface.co/cognitivecomputations). The purpose of this dataset is to train R1-style reasoning models.

This is a reformatted version of the Flash subset for ease of use. It adds the model's response to the conversation with the following special tokens: `<|begin_of_thought|>`, `<|end_of_thought|>`, `<|begin_of_solution|>`, `<|end_of_solution|>`.

Please like the [original dataset](https://huggingface.co/datasets/cognitivecomputations/dolphin-r1) if you enjoy this reformatted version. Thanks to Eric Hartford for this!

# 海豚R1 Flash 🐬

本数据集采用Apache-2.0开源协议，由Eric Hartford（https://huggingface.co/ehartford）与Cognitive Computations（https://huggingface.co/cognitivecomputations）共同整理构建。本数据集的核心目标是用于训练R1风格的推理模型。

本版本为Flash子集的重排版优化版本，旨在提升使用便捷性。本次重排版为对话数据新增了模型回复内容，并通过以下特殊Token进行标记：`<|begin_of_thought|>`, `<|end_of_thought|>`, `<|begin_of_solution|>`, `<|end_of_solution|>`。

若您喜爱本次重排版本，欢迎为原始数据集（https://huggingface.co/datasets/cognitivecomputations/dolphin-r1）点赞。在此衷心感谢Eric Hartford的贡献！