mpasila/Finnish-ShareGPT-Small-V2

---
license: mit
language:
- fi
size_categories:
- 1K<n<10K
---
I added some new instructions that use the same topics as the FIN-Bench but those are just the topics which I still generate using the same prompt I used for generating the rest of the instructions. (also added some other unrelated instructions and removed/replaced some)

Another change is format. This now uses ShareGPT formatting.

I'll make a new dataset after this using a different prompt to get it a bit more conversational.

```
From the previous dataset: This is a larger version of my previous dataset containing the same data but now with a lot more instructions generated using GPT-4-Turbo and GPT-4o. It may have issues that I've not noticed.

It contains parts of the Stanford Alpaca seed tasks translated to Finnish using DeepL/ChatGPT with manual edits to fix any issues. It also contains some ChatGPT generated instructions.
```

原始信息汇总

数据集概述

基本信息

• 许可证: MIT
• 语言: 芬兰语 (fi)
• 大小类别: 1K<n<10K

内容描述

• 数据集包含使用GPT-4-Turbo和GPT-4o生成的指令，这些指令基于FIN-Bench的主题，并进行了部分添加、移除或替换。
• 数据集还包含斯坦福Alpaca种子任务的芬兰语翻译，这些翻译通过DeepL/ChatGPT生成，并经过手动编辑以修正任何问题。
• 此外，数据集中还包含一些由ChatGPT生成的指令。

格式变更

• 数据集现在采用ShareGPT格式。