Opus-WritingPrompts

## Opus Writing Prompts
This is a dataset containing 3008 short stories, generated by an unrestrained Claude Opus using Reddit's Writing Prompts as a source. Each sample is generally between 4000-6000 characters long.

These stories were thoroughly cleaned and then further enriched with a title and a series of applicable genres.

**Disclaimer:** This dataset is extremely varied and includes erotica. You have been warned.

**Three files are included:**
- A ShareGPT dataset, ready to be used for training.
- A "raw" version, containing prompt/title/genre/story columns.

- **2024-05-28:** I've included a GPT 3.5 generated version in a similar format as the raw file, meant to act as a negative counterpart for KTO training purposes.

- **2024-06-02:** Another Opus generated version was added, using the same writing prompts. These have been checked and all are distinct enough to be used as an expansion for the existing set.

If you find any errors, please let me know!

**Opus写作提示数据集（Opus Writing Prompts）**
本数据集共收录3008篇短篇小说，均由无约束版本的Claude Opus模型以Reddit平台的写作提示（Writing Prompts）板块为素材生成。单条样本的字符长度通常介于4000至6000之间。

所有故事均经过全面的清洗预处理，并额外补充了标题与多组合适的体裁标签。

**免责声明：** 本数据集内容跨度极广，包含色情文学相关内容，敬请使用者知悉。

**包含文件与更新：**
- 一份ShareGPT格式数据集，可直接用于模型训练。
- 一份“原始版”文件，包含提示词、标题、体裁、故事四列数据。
- **2024年5月28日更新：** 新增GPT 3.5生成的同格式版本，可作为负向对照样本用于KTO训练。
- **2024年6月2日更新：** 新增另一版由Opus生成的故事，仍基于相同的写作提示源。经核查，所有故事均具备足够差异性，可作为现有数据集的扩充版本。

若发现任何错误，欢迎反馈指正！