datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
CulturaY
CulturaY: A Large Cleaned Multilingual Dataset of 75 Languages
Dataset Summary
From the team that brought you CulturaX, we present CulturaY, another substantial multilingual dataset of 15TB (uncompressed)/3TB (zstd-compressed) that applies the same dataset cleaning methodology to the HPLT v1.1 dataset.
Please note that HPLT v1.2 has also been released and is an alternative verison with different cleaning methodolgies.
This data was used in part to train our SOTA… See the full description on the dataset page: https://huggingface.co/datasets/Viet-Mistral/CulturaY.based-chat-v0.1-Mistral-Nemo-Base-2407
Based-Chat v0.1 (Mistral Nemo Base 2407)
This dataset was developed as part of an exploration into understanding the necessity of supervised datasets for fine-tuning base LLMs into conversational models.
It's a synthetic dataset created with Mistral-Nemo-Base-2407, and used to fine-tune that model, producing relay-v0.1-Mistral-Nemo-2407.
Methodology
This synthetic dataset is generated using the following as conversation starters:
facebook/empathetic_dialogues… See the full description on the dataset page: https://huggingface.co/datasets/danlou/based-chat-v0.1-Mistral-Nemo-Base-2407.robuchan-data
Robuchan Dataset
Synthetic dietary recipe adaptation dataset for fine-tuning language models. Each example is a chat-format conversation where a user provides a recipe and dietary restriction, and the assistant produces a structured adaptation.
Generated for the Mistral AI Worldwide Hackathon Tokyo (Feb 28 - Mar 1, 2026).
Associated model: sumitdotml/robuchan
Dataset Structure
Splits
Split
Rows
Purpose
train
1,090
Fine-tuning training set… See the full description on the dataset page: https://huggingface.co/datasets/mistral-hackaton-2026/robuchan-data.mistral-legal-french-dataset
Mistral Legal French Dataset
A fine-tuning dataset for French legal domain, optimized with curriculum learning strategy.
📋 Table of Contents
Overview
Dataset Composition
Methodology
1. Chain-of-Thought Generation
2. LegalKit Extraction
3. Curriculum Learning Fusion
Data Format
Quality Metrics
Usage
Citations
License
🎯 Overview
This dataset was created to fine-tune Mistral-7B-Instruct-v0.3 on French legal domain tasks. It combines two… See the full description on the dataset page: https://huggingface.co/datasets/davidpistori/mistral-legal-french-dataset.mistral-legal-french-dataset
Mistral Legal French Dataset
A fine-tuning dataset for French legal domain, optimized with curriculum learning strategy.
📋 Table of Contents
Overview
Dataset Composition
Methodology
1. Chain-of-Thought Generation
2. LegalKit Extraction
3. Curriculum Learning Fusion
Data Format
Quality Metrics
Usage
Citations
License
🎯 Overview
This dataset was created to fine-tune Mistral-7B-Instruct-v0.3 on French legal domain tasks. It combines two complementary… See the full description on the dataset page: https://huggingface.co/datasets/VinceGx33/mistral-legal-french-dataset.Medical-QA-Mistral7B-FinetuningZero_SFT_Ja_by_Mistral_Small
DataPilot/Zero_SFT_Ja_by_Mistral_Small
このデータセットは、日本語で記述された高品質な合成プロンプトとそのAI出力を収録しています。すべてのデータは Mistral Small 3.1 24B Instruct 2503 モデルを使用してゼロから合成されています。
概要
項目
詳細
データセット名
DataPilot/Zero_SFT_Ja_by_Mistral_Small
言語
日本語
データ作成方法
完全自動生成（モデルによるゼロショット合成）
使用モデル
Mistral Small 3.1 24B Instruct 2503
フォーマット
JSONL（id, input, output, conversation）
ライセンス
Apache-2.0
作成コード
foxn2000/zero_one_instruction
データセット構造
データセットには以下のカラムが含まれています。
カラム名… See the full description on the dataset page: https://huggingface.co/datasets/DataPilot/Zero_SFT_Ja_by_Mistral_Small.JMT-Bench-result_self-rewarding_Mistral-7B-lora
JMT-Bench result
Answer language
JMT-Benchの回答のうち、Englishで回答した件数
Model
Count
mistralai/Mistral-7B-v0.3
25
HachiML/Mistral-7B-v0.3-m1-lora
7
HachiML/Mistral-7B-v0.3-m2-lora
7
HachiML/Mistral-7B-v0.3-m3-lora
2
Cybersecurity_Reasoning_Dataset_MistralFamily_7b
Cybersecurity Reasoning Dataset (v6.0)
A high-fidelity, forensics-mapped training corpus for cybersecurity reasoning and SOC
automation — formatted for the Mistral / Llama instruct family.
Format-specific dataset. Every record uses the Alpaca-style
### Instruction: / ### Response: template native to Mistral/Llama instruct models.
A model-agnostic version of this corpus (Mistral, DeepSeek, ChatML, and Gemma
variants rendered from one neutral canonical source) is published… See the full description on the dataset page: https://huggingface.co/datasets/dpevzner/Cybersecurity_Reasoning_Dataset_MistralFamily_7b.