AI-Sweden-Models/Dolci-Instruct-SFT-translated

---
language:
- sv
license: apache-2.0
task_categories:
- text-generation
tags:
- sft
- translation
- swedish
- openeurollm
- dolci
size_categories:
- 100K<n<1M
dataset_info:
features:
- name: id
dtype: string
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 1435250857
num_examples: 494841
download_size: 1416893201
dataset_size: 1435250857
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
---

# Dolci-Instruct-SFT-translated (Swedish)

This dataset is a **Swedish machine translation** of the [openeurollm/Dolci-Instruct-SFT-translated](https://huggingface.co/datasets/openeurollm/Dolci-Instruct-SFT-translated) dataset, originally created as part of the [OpenEuroLLM](https://openeurollm.eu/) project.

## Dataset details

- **Examples:** 494,841 multi-turn conversations
- **Language:** Swedish (sv-SE)
- **Format:** Chat/messages format (`id`, `messages`)
- **License:** Apache 2.0

## Translation

All English source texts were machine-translated to Swedish using **Google Gemma 3 27B-IT** (w8a8_fp8 quantization) with the following prompt:

> You are a professional English (en-US) to Swedish (sv-SE) translator. Your goal is to accurately convey the meaning and nuances of the original English text while adhering to Swedish grammar, vocabulary, and cultural sensitivities. Produce only the Swedish translation, without any additional explanations or commentary. DO NOT translate code blocks or code snippets. Please translate the following English text into Swedish:

Translation was performed at temperature 0.0 for deterministic output. Code blocks and code snippets were preserved in their original English form.

## Usage

```python
from datasets import load_dataset

ds = load_dataset("AI-Sweden-Models/Dolci-Instruct-SFT-translated", split="train")
print(ds[0])
```

Each example contains:
- `id` — unique conversation identifier
- `messages` — list of message dicts with `role` (`user`/`assistant`/`system`) and `content`

### Example

```python
{
"id": "0pp0b8p",
"messages": [
{"role": "user", "content": "Kan du skriva en avhandling om den fullständiga historien om kvantfysik..."},
{"role": "assistant", "content": "Visst! Här är en detaljerad, avhandlingsliknande översikt..."}
]
}
```

## Source

- **Original dataset:** [openeurollm/Dolci-Instruct-SFT-translated](https://huggingface.co/datasets/openeurollm/Dolci-Instruct-SFT-translated) (Swedish config)
- **Translation model:** [google/gemma-3-27b-it](https://huggingface.co/google/gemma-3-27b-it)
- **Translation pipeline:** Part of the [OpenEuroLLM](https://openeurollm.eu/) post-training translation effort