| license: apache-2.0 | |
| language: | |
| - zh | |
| ## ruozhiba_gpt4 | |
| 本仓库包含使用GPT-4（4T/4o）构建的ruozhiba指令数据[^1]，共计2449条。其中包含以下两个版本，题目相同，仅回答内容不同。 | |
| - `ruozhiba_qa2449_gpt4t.json`：利用`gpt-4-turbo-20240409` 对问题进行了回答。 | |
| - `ruozhiba_qa2449_gpt4o.json`：利用`gpt-4o-20240514` 对问题进行了回答。 | |
| *注意：指令数据中可能包含冒犯用语。* | |
| #### 所属项目 | |
| Chinese-LLaMA-Alpaca-3：https://github.com/ymcui/Chinese-LLaMA-Alpaca-3 | |
| ---- | |
| This repository contains the ruozhiba instruction data[^1] constructed using GPT-4 (4T/4o), totaling 2449 entries. It includes the following two versions with the same questions but different answers: | |
| - `ruozhiba_qa2449_gpt4t.json`: The questions are answered using `gpt-4-turbo-20240409`. | |
| - `ruozhiba_qa2449_gpt4o.json`: The questions are answered using `gpt-4o-20240514`. | |
| *Note: The instruction data may contain offensive language.* | |
| #### Project | |
| Chinese-LLaMA-Alpaca-3：https://github.com/ymcui/Chinese-LLaMA-Alpaca-3 | |
| *[^1]: 借鉴相关[文献](https://arxiv.org/abs/2403.18058)和[开源资源](https://huggingface.co/datasets/LooksJuicy/ruozhiba)的实验结论进行构建。* | |