toroe/Soofi-Think-SFT-V2-firsthalf-FR

---
language:
- fr
license: other
task_categories:
- text-generation
task_ids:
- language-modeling
tags:
- reasoning
- thinking
- chain-of-thought
- sft
- translation
- french
- math
- science
- code
- chat
- tool-calling
pretty_name: Soofi-Think-SFT-V2-firsthalf-FR
size_categories:
- 1M<n<10M
---

# Soofi-Think-SFT-V2-firsthalf-FR

French-translated version of [toroe/Soofi-Think-SFT-V2-firsthalf](https://huggingface.co/datasets/toroe/Soofi-Think-SFT-V2-firsthalf) — a large-scale supervised fine-tuning dataset featuring chain-of-thought reasoning traces (`<think>...</think>`) across math, science, code, tool-calling, and general instruction-following tasks.

The translation was produced using **Qwen3-32B** via [vLLM](https://github.com/vllm-project/vllm), applying professional-grade translation prompts targeting standard French suitable for international francophone audiences.

---

## Dataset Summary

| Property | Value |
|---|---|
| Language | French (`fr`) |
| Source dataset | `toroe/Soofi-Think-SFT-V2-firsthalf` |
| Total rows | ~2.37M |
| Translation model | `Qwen/Qwen3-32B` (FP8 quantization) |
| Format | Chat-style JSONL (`messages` field) |
| Thinking traces | Preserved with `<think>…</think>` tags |

---

## Source Datasets

The rows in this dataset originate from a broad blend of high-quality English SFT corpora. The `dataset_name` and `source` fields identify the provenance of each row. Known source collections include:

- **Dolci-Think-SFT-7B** — OpenThoughts3 (math, science, code), WildJailbreak R1, WildChat R1, WildGuardMix R1, Aya-100k R1, Persona-precise-IF R1, SYNTHETIC-2-SFT, Nemotron-post-training subset, correct-python-sft
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2 (general)** — SlimOrca, HuggingFaceTB/smoltalk, mmlu_auxiliary_train, ShareGPT_Vicuna_unfiltered, GPTeacher-General-Instruct, flan_v2, synthetic, nvidia/Nemotron-Post-Training-Dataset-v1
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2 (math)** — NuminaMath-CoT, OpenMathReasoning
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2 (science)** — Nemotron-Post-Training-Dataset-v1-stem, synthetic
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2 (code)** — OpenCodeReasoning, leetcode
- **Nemotron-Cascade-SFT-Stage-2 (tool-calling)** — Nemotron-Post-Training-Dataset-v1-tool-calling
- **Nemotron-Science-v1** — MCQ, RQA
- **Llama-Nemotron-Post-Training-Dataset** — Science subset
- **Nemotron-Instruction-Following-Chat-v1** — nemotron_v3_chat

---

## Dataset Structure

Each row is a JSONL record with the following fields:

```json
{
"row_index": 0,
"dataset_name": "Dolci-Think-SFT-7B",
"source": "saumyamalik/OpenThoughts3-full-filtered-science-decontam-v2",
"ds_uid": 839609,
"language": "french",
"messages": [
{"role": "user", "content": "..."},
{"role": "assistant", "content": "<think>\n...\n</think>\n..."}
],
"thinking_chunked": null
}
```

### Fields

| Field | Type | Description |
|---|---|---|
| `row_index` | int | Original row index in the source dataset |
| `dataset_name` | string | High-level source collection name |
| `source` | string | Specific upstream HuggingFace dataset/split |
| `ds_uid` | int | Unique ID from the source dataset |
| `language` | string | Always `"french"` |
| `messages` | list | Chat-format turns: `system` / `user` / `assistant` |
| `thinking_chunked` | bool or null | `true` if the `<think>` block was too long to translate in one pass and was split into chunks |

### Message Format

Assistant turns that include a reasoning trace are formatted as:

```
<think>
[translated reasoning trace]
</think>
[translated final answer]
```

Tool-calling rows may include a `system` turn with function signatures, which are intentionally left in English as they contain code-like structured content (function names, JSON schemas, identifiers).

---

## Translation Methodology

Translation was performed with a custom vLLM-based pipeline. Key design decisions:

- **Model:** `Qwen/Qwen3-32B` with FP8 weight quantization and prefix caching enabled
- **Decoding:** Near-greedy sampling (temperature `0.1`, top-p `1.0`) for translation stability
- **Context management:** Input token budgets are computed as `usable / (1 + output_ratio)` where `output_ratio=1.1`, ensuring sufficient room for output generation
- **Long-text chunking:** Fields exceeding the token limit are split on paragraph boundaries (falling back to line, then word boundaries) and translated in chunks, then reassembled. Rows where this occurred are flagged with `thinking_chunked: true`
- **Batch efficiency:** All non-chunked fields across a batch are sent to vLLM in a single call; chunked fields are also batched together in large calls to maximize throughput
- **Register:** Standard French with appropriate formality, suitable for international francophone audiences
- **Preserved elements:** Code, variable names, LaTeX/mathematical notation, file paths, URLs, tool/function signatures, and quoted literals are left in English

### Translation Prompt Guidelines (summary)

The system prompt instructed the model to:

1. Output **only** the translated text — no meta-commentary or explanations
2. Translate **all** natural-language prose; leave code, identifiers, and literals unchanged
3. Preserve formatting, tone, and formality level of the original
4. Adapt cultural references appropriately for French-speaking audiences
5. Maintain consistent terminology throughout each document

---

## Intended Uses

This dataset is intended for:

- **Multilingual SFT / instruction tuning** of language models targeting French-speaking users
- **Cross-lingual reasoning** research (chain-of-thought in French)
- **Distillation** of reasoning capabilities into smaller French-language models
- **Tool-use and function-calling** training in a French context
- Benchmarking **translation quality** of reasoning-heavy content

---

## Limitations

- Translations are machine-generated and may contain errors, particularly for highly domain-specific or ambiguous content
- Very long reasoning traces that required chunked translation (`thinking_chunked: true`) may have minor coherence issues at chunk boundaries
- Tool-calling `system` prompts are intentionally kept in English, as they contain structured technical content (JSON schemas, function signatures) that must remain machine-readable
- Technical terms and proper nouns are generally preserved in English, which reflects standard practice for French technical writing but may not suit all use cases
- The dataset inherits any biases, errors, or quality issues present in the original English source datasets

---

## Citation

If you use this dataset, please also cite the original upstream datasets and the Qwen3 model used for translation.

```bibtex
@misc{soofi-think-sft-v2-fr,
title = {Soofi-Think-SFT-V2-firsthalf-FR},
author = {toroe},
year = {2025},
howpublished = {\url{https://huggingface.co/datasets/toroe/Soofi-Think-SFT-V2-firsthalf-FR}},
note = {French translation of Soofi-Think-SFT-V2-firsthalf using Qwen3-32B via vLLM}
}
```

language:
- 法语（fr）
许可：其他
任务类别：
- 文本生成
任务子类型：
- 语言建模
标签：
- 推理
- 思考
- 思维链（chain-of-thought）
- 监督微调（SFT, Supervised Fine-Tuning）
- 翻译
- 法语
- 数学
- 科学
- 代码
- 聊天
- 工具调用
美观名称：Soofi-Think-SFT-V2-firsthalf-FR
规模类别：
- 100万 < 样本量 < 1000万

# Soofi-Think-SFT-V2-firsthalf-FR

本数据集为[toroe/Soofi-Think-SFT-V2-firsthalf](https://huggingface.co/datasets/toroe/Soofi-Think-SFT-V2-firsthalf)的法语翻译版本，后者是大规模监督微调数据集，涵盖数学、科学、代码、工具调用及通用指令跟随任务中的思维链（chain-of-thought）推理轨迹（格式为`<think>...</think>`）。

本次翻译通过**Qwen3-32B**借助[vLLM](https://github.com/vllm-project/vllm)完成，采用面向国际法语使用者的标准法语专业级翻译提示词。

---

## 数据集摘要

| 属性 | 值 |
|---|---|
| 语言 | 法语（`fr`） |
| 源数据集 | `toroe/Soofi-Think-SFT-V2-firsthalf` |
| 总样本量 | 约237万 |
| 翻译模型 | `Qwen/Qwen3-32B`（FP8量化） |
| 格式 | 聊天式JSONL（含`messages`字段） |
| 推理轨迹 | 保留`<think>…</think>`标签 |

---

## 源数据集

本数据集的样本源自多个高质量英语监督微调语料的广泛混合。`dataset_name`与`source`字段用于标识每条样本的具体来源。已知的源数据集集合包括：
- **Dolci-Think-SFT-7B** — OpenThoughts3（数学、科学、代码）、WildJailbreak R1、WildChat R1、WildGuardMix R1、Aya-100k R1、Persona-precise-IF R1、SYNTHETIC-2-SFT、Nemotron后训练子集、correct-python-sft
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2（通用场景）** — SlimOrca、HuggingFaceTB/smoltalk、mmlu_auxiliary_train、ShareGPT_Vicuna_unfiltered、GPTeacher-General-Instruct、flan_v2、synthetic、nvidia/Nemotron-Post-Training-Dataset-v1
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2（数学场景）** — NuminaMath-CoT、OpenMathReasoning
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2（科学场景）** — Nemotron-Post-Training-Dataset-v1-stem、synthetic
- **Nemotron-Cascade-SFT-Stage-1 / Stage-2（代码场景）** — OpenCodeReasoning、leetcode
- **Nemotron-Cascade-SFT-Stage-2（工具调用场景）** — Nemotron-Post-Training-Dataset-v1-tool-calling
- **Nemotron-Science-v1** — MCQ、RQA
- **Llama-Nemotron-Post-Training-Dataset** — 科学子集
- **Nemotron-Instruction-Following-Chat-v1** — nemotron_v3_chat

---

## 数据集结构

每条样本为一条JSONL格式记录，包含以下字段：
json
{
"row_index": 0,
"dataset_name": "Dolci-Think-SFT-7B",
"source": "saumyamalik/OpenThoughts3-full-filtered-science-decontam-v2",
"ds_uid": 839609,
"language": "french",
"messages": [
{"role": "user", "content": "..."},
{"role": "assistant", "content": "<think>
...
</think>
..."}
],
"thinking_chunked": null
}

### 字段说明
| 字段 | 类型 | 描述 |
|---|---|---|
| `row_index` | 整数 | 源数据集中的原始样本索引 |
| `dataset_name` | 字符串 | 高层级源数据集集合名称 |
| `source` | 字符串 | 具体上游Hugging Face数据集/拆分 |
| `ds_uid` | 整数 | 源数据集的唯一标识符 |
| `language` | 字符串 | 固定为"french" |
| `messages` | 列表 | 聊天格式对话轮次：包含`system`/`user`/`assistant`角色 |
| `thinking_chunked` | 布尔值或空值 | 若`<think>`块过长无法一次性翻译并拆分为多个块，则为`true` |

### 对话格式
包含推理轨迹的助手回复格式如下：

<think>
[翻译后的推理轨迹]
</think>
[翻译后的最终答案]

工具调用样本可能包含带有函数签名的`system`轮次，此类内容有意保留为英语，因其包含类代码的结构化内容（函数名、JSON Schema、标识符）。

---

## 翻译方法

翻译通过基于vLLM的自定义流水线完成，关键设计决策如下：
- **模型**：`Qwen/Qwen3-32B`，启用FP8权重量化与前缀缓存功能
- **解码策略**：近似贪心采样（温度`0.1`，top-p`1.0`）以保障翻译稳定性
- **上下文管理**：输入令牌预算按`可用长度 / (1 + 输出比例)`计算，其中`输出比例=1.1`，确保有足够空间生成输出内容
- **长文本分块**：超出令牌限制的字段将按段落边界拆分（优先按段落，回退至行边界，最终回退至词边界），分块翻译后重新拼接。出现该情况的样本将被标记为`thinking_chunked: true`
- **批量效率**：批次内所有未分块的字段将通过单次vLLM调用完成翻译；分块字段也将批量处理以最大化吞吐量
- **语域规范**：采用标准法语，语气正式，适配国际法语使用者群体
- **保留元素**：代码、变量名、LaTeX/数学符号、文件路径、URL、工具/函数签名以及引用文本均保留为英语

### 翻译提示词指南（摘要）
系统提示词要求模型：
1. 仅输出翻译后的文本，不得添加任何元注释或解释性内容
2. 翻译所有自然语言文本；代码、标识符与引用文本保持原样
3. 保留原文的格式、语气与正式程度
4. 针对法语使用者群体适配文化参考内容
5. 在整个文档中保持术语一致性

---

## 预期用途

本数据集适用于：
- 面向法语使用者的大语言模型（LLM/Large Language Model）多语言监督微调/指令调优
- 跨语言推理研究（法语环境下的思维链推理）
- 将推理能力蒸馏为小型法语语言模型
- 法语语境下的工具使用与函数调用训练
- 推理密集型内容的翻译质量基准测试

---

## 局限性

- 翻译为机器生成，可能存在错误，尤其是针对高度领域特定或歧义的内容
- 需要分块翻译的超长推理轨迹（`thinking_chunked: true`）在分块边界处可能存在轻微的连贯性问题
- 工具调用的`system`提示词有意保留为英语，因其包含结构化技术内容（JSON Schema、函数签名），需保持机器可读性
- 技术术语与专有名词通常保留为英语，这符合法语技术写作的标准惯例，但可能不适用于所有使用场景
- 本数据集继承了原始英语源数据集中存在的所有偏见、错误或质量问题

---

## 引用说明

若使用本数据集，请同时引用原始上游数据集以及用于翻译的Qwen3模型。
bibtex
@misc{soofi-think-sft-v2-fr,
title = {Soofi-Think-SFT-V2-firsthalf-FR},
author = {toroe},
year = {2025},
howpublished = {url{https://huggingface.co/datasets/toroe/Soofi-Think-SFT-V2-firsthalf-FR}},
note = {基于Qwen3-32B通过vLLM完成的Soofi-Think-SFT-V2-firsthalf法语翻译}
}