datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
chi-bench
Clinical Healthcare In-Situ Environment
Task fixtures for a long-horizon, policy-rich healthcare-workflow agent benchmark
What is in this dataset
CHI-Bench evaluates AI agents on end-to-end U.S. healthcare workflows across three long-horizon domains: provider prior authorization, payer utilization management, and population care management. Each task hands the agent a clinical case in a high-fidelity simulator of 20 healthcare apps exposed over MCP, with a 1… See the full description on the dataset page: https://huggingface.co/datasets/actava/chi-bench.fineweb-edu
📚 FineWeb-Edu
1.3 trillion tokens of the finest educational data the 🌐 web has to offer
Paper: https://arxiv.org/abs/2406.17557
What is it?
📚 FineWeb-Edu dataset consists of 1.3T tokens and 5.4T tokens (FineWeb-Edu-score-2) of educational web pages filtered from 🍷 FineWeb dataset. This is the 1.3 trillion version.
To enhance FineWeb's quality, we developed an educational quality classifier using annotations generated by LLama3-70B-Instruct. We… See the full description on the dataset page: https://huggingface.co/datasets/chilomax/fineweb-edu.chinese-fineweb-edu-v2
This version is deprecated. We recommend you to use the newest version Fineweb-edu-chinese-v2.1 !
Chinese Fineweb Edu Dataset V2 [中文] [English]
[OpenCSG Community] [👾github] [wechat] [Twitter]
📖Technical Report
Chinese Fineweb Edu Dataset V2 is a comprehensive upgrade of the original Chinese Fineweb Edu, designed and optimized for natural language processing (NLP) tasks in the education sector. This high-quality Chinese pretraining dataset has… See the full description on the dataset page: https://huggingface.co/datasets/opencsg/chinese-fineweb-edu-v2.smoltalk-chinese
Chinese SmolTalk Dataset [中文] [English]
[OpenCSG Community] [👾github] [wechat] [Twitter]
📖Technical Report
smoltalk-chinese is a Chinese fine-tuning dataset constructed with reference to the SmolTalk dataset. It aims to provide high-quality synthetic data support for training large language models (LLMs). The dataset consists entirely of synthetic data, comprising over 700,000 entries. It is specifically designed to enhance the performance of Chinese… See the full description on the dataset page: https://huggingface.co/datasets/opencsg/smoltalk-chinese.ChinaTravel
ChinaTravel Query Dataset
This dataset is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).
ChinaTravel is an open-ended travel-planning benchmark with compositional
constraint validation for language agents. See the
paper,
Hugging Face paper page,
code, and
bilingual sandbox database
(ModelScope mirror)
for the complete benchmark resources.
Introduction
For a given query, a language agent uses the sandbox tools to collect
information and… See the full description on the dataset page: https://huggingface.co/datasets/LAMDA-NeSy/ChinaTravel.Chinese-DeepSeek-R1-Distill-data-110k
中文基于满血DeepSeek-R1蒸馏数据集（Chinese-Data-Distill-From-R1）
🤗 Hugging Face | 🤖 ModelScope | 🚀 Github | 📑 Blog
注意：提供了直接SFT使用的版本，点击下载。将数据中的思考和答案整合成output字段，大部分SFT代码框架均可直接直接加载训练。
本数据集为中文开源蒸馏满血R1的数据集，数据集中不仅包含math数据，还包括大量的通用类型数据，总数量为110K。
为什么开源这个数据？
R1的效果十分强大，并且基于R1蒸馏数据SFT的小模型也展现出了强大的效果，但检索发现，大部分开源的R1蒸馏数据集均为英文数据集。 同时，R1的报告中展示，蒸馏模型中同时也使用了部分通用场景数据集。
为了帮助大家更好地复现R1蒸馏模型的效果，特此开源中文数据集。
该中文数据集中的数据分布如下：… See the full description on the dataset page: https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k.China-K12-STEM-10K-CoT-Reasoning
K12-STEM-CoT-Chinese
1.54M Chinese K12 STEM problems with chain-of-thought solutions, 48% with diagrams.
The largest structured Chinese math/physics/chemistry reasoning dataset.
This is a curated sample (10,000 problems) of the full 1.54M dataset available via API.
Full Dataset Access
Access the full 1,540,000+ problems via API →
This Sample
Full API
Total problems
10,025
1,540,000+
With CoT solutions
10,025
1,490,000+
With diagrams
6,093
740,000+… See the full description on the dataset page: https://huggingface.co/datasets/lfaviate/China-K12-STEM-10K-CoT-Reasoning.Chinese_Debate_Documents
Dataset Card for Chinese Debate Documents
ASR-transcribed corpus of competitive Mandarin Chinese university debates,
speaker-segmented and timestamped, with topic / round / team metadata parsed
from the source filenames.
Loading the Dataset
from datasets import load_dataset
ds = load_dataset("DBWBD/Chinese_Debate_Documents", split="train")
print(ds[0]["topic"], "—", ds[0]["team_a"], "vs", ds[0]["team_b"])
for seg in ds[0]["segments"][:3]:
print(f"… See the full description on the dataset page: https://huggingface.co/datasets/DBWBD/Chinese_Debate_Documents.china-effective-laws-regulations
全国现行法律法规合集
现行有效的中华人民共和国法律、行政法规、监察法规、地方性法规、司法解释结构化文本。一部法规一行，一条法条一行，供查阅、检索、RAG 和法律 NLP 使用。
数据来自全国人大常委会办公厅 国家法律法规数据库，下载口径为官网的 「有效及尚未生效」。正文由 Word 原文用脚本抽取，未经大模型改写。
这不是官方汇编，不能替代公报或标准文本，也不能作为法律意见。 电子文本与标准文本不一致时，以法律规定的标准文本为准。
快照日期：2026-08-26
效力说明
本数据集 以现行有效法律法规为主体：
效力 status
法规份数
说明
有效
17,649
现行有效，默认应使用这一部分
尚未生效
7
已公布、施行日晚于快照日
失效
45
文件名含「失效」，多为已到期的全国人大常委会试点授权决定
使用时请筛选 status == "有效"，即可得到现行有效文本。同一部法若有修正前后多个版本，均予保留，用 filename_date 区分，采用最新日期即可。… See the full description on the dataset page: https://huggingface.co/datasets/senry5433/china-effective-laws-regulations.Chinese-DeepSeek-R1-Distill-data-110k-SFT
中文基于满血DeepSeek-R1蒸馏数据集（Chinese-Data-Distill-From-R1）
🤗 Hugging Face | 🤖 ModelScope | 🚀 Github | 📑 Blog
注意：该版本为，可以直接SFT使用的版本，将原始数据中的思考和答案整合成output字段，大部分SFT代码框架均可直接直接加载训练。
本数据集为中文开源蒸馏满血R1的数据集，数据集中不仅包含math数据，还包括大量的通用类型数据，总数量为110K。
为什么开源这个数据？
R1的效果十分强大，并且基于R1蒸馏数据SFT的小模型也展现出了强大的效果，但检索发现，大部分开源的R1蒸馏数据集均为英文数据集。 同时，R1的报告中展示，蒸馏模型中同时也使用了部分通用场景数据集。
为了帮助大家更好地复现R1蒸馏模型的效果，特此开源中文数据集。该中文数据集中的数据分布如下：
Math：共计36568个样本，
Exam：共计2432个样本，
STEM：共计12648个样本，… See the full description on the dataset page: https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT.TC260-Chinese-Safety-Prompts
TC260 Chinese Safety Prompts V1
Public research dataset containing synthetic Chinese safety-testing prompts.
Records have different quality tiers; the full dataset must not be described
as human-verified or Gold data.
这是一个面向中文生成式人工智能安全评测研究的合成测试提示数据集。候选数据
由项目冻结的 tc260-generator-v3.2 生成，并经过结构校验、凭据与内部路径
扫描、精确去重和四字shingle近似去重。
本数据集不是TC260或任何国家标准机构发布、认可或认证的官方数据集。
类别名称和映射用于研究性实现，不构成法律、监管或合规结论。
数据规模
原始生成规模：5,000条候选；结构清洗后正式发布4,997条（剔除2条标记泄漏和1条重复记录）。
A.1至A.4：4… See the full description on the dataset page: https://huggingface.co/datasets/BBBBBBBBBBBQ/TC260-Chinese-Safety-Prompts.Chinese_interactive_novels_3k
中文互动小说结构化语料
This dataset contains uncleaned (!) 3534 structured Chinese interactive novels (中文互动小说), accounting for around 0.25B (gpt-3.5) tokens in total.
All contents are parsed from certain online sources.
Usage
This dataset can be potentially used for LLM training. But be aware that you'd better clean the data yourself to remove undesired low-quality contents.
Each novel is a dict structured as follows:
class Novel:
book_title: str
book_author: str… See the full description on the dataset page: https://huggingface.co/datasets/mrzjy/Chinese_interactive_novels_3k.ielts-writing-task2-essays
📚 IELTS Writing Task 2 Essays & Feedback Dataset (Writing9)
Dataset Summary
This dataset contains 8,000+ real IELTS Writing Task 2 essays crawled from Writing9. It covers 128 real IELTS exam questions categorized into 25 topics (such as Art, Business, Education, Technology, Environment, Government, Health, etc.).
Each record includes:
essay_id: Unique identifier on Writing9
topic: Topic category (e.g. Art, Business and Companies, Cities)
question: Cleaned IELTS… See the full description on the dataset page: https://huggingface.co/datasets/chillies/ielts-writing-task2-essays.smoltalk-chinese
Chinese SmolTalk Dataset [中文] [English]
[OpenCSG Community] [👾github] [wechat] [Twitter]
📖Technical Report
smoltalk-chinese is a Chinese fine-tuning dataset constructed with reference to the SmolTalk dataset. It aims to provide high-quality synthetic data support for training large language models (LLMs). The dataset consists entirely of synthetic data, comprising over 700,000 entries. It is specifically designed to enhance the performance of Chinese… See the full description on the dataset page: https://huggingface.co/datasets/sunorme/smoltalk-chinese.chinese-writing-bench-judgements-gpt-5.4
Zhiyin: Exploring the Frontier of Chinese LLM Writing
Website • GitHub • Hugging Face
Zhiyin is an LLM-as-a-judge benchmark for Chinese writing evaluation. This V1 release features 280 test cases across 18 diverse writing tasks.
Benchmark Overview
Our evaluation method relies on pairwise comparison. A powerful language model (O3) acts as the judge, scoring a model's response relative to a fixed baseline (GPT-4.1), which is anchored at a score of 5.
Scoring… See the full description on the dataset page: https://huggingface.co/datasets/zake7749/chinese-writing-bench-judgements-gpt-5.4.chi-bench
Clinical Healthcare In-Situ Environment
Task fixtures for a long-horizon, policy-rich healthcare-workflow agent benchmark
What is in this dataset
CHI-Bench evaluates AI agents on end-to-end U.S. healthcare workflows across three long-horizon domains: provider prior authorization, payer utilization management, and population care management. Each task hands the agent a clinical case in a high-fidelity simulator of 20 healthcare apps exposed over MCP, with a 1… See the full description on the dataset page: https://huggingface.co/datasets/test-alexpouliquen/chi-bench.smoltalk-chinese-QwQ-Distrill
smoltalk-chinese-QwQ-Distrill [中文] [English]
📖Technical Report
smoltalk-chinese-QwQ-Distrill is a Chinese fine-tuning dataset constructed with reference to the SmolTalk-Chinese dataset. It aims to provide high-quality synthetic reasoning data support for training large language models (LLMs). The dataset consists entirely of synthetic data, comprising over 700,000 entries. It is specifically designed to enhance the performance of Chinese LLMs across various tasks… See the full description on the dataset page: https://huggingface.co/datasets/ChinaunicomSoftware/smoltalk-chinese-QwQ-Distrill.chinese_novel
Space Grimoire Novel Corpus (Traditional Chinese)
Full text of the Traditional Chinese web novel 空間魔導書與少年魔法師 (The Space Grimoire and the Young Mage), split by chapter.
Field
Value
Author
睡半夜怎麼三更
License
CC BY 4.0
Language
Traditional Chinese (zh-Hant-TW)
Genre
Fantasy, steampunk, political intrigue
Chapters
283 (interludes included)
Parts
10
Paragraphs
30,760
Characters (body text)
1,737,806
Version
2026-09-13
Companion dataset… See the full description on the dataset page: https://huggingface.co/datasets/asd567557275/chinese_novel.greathangpt-classical-chinese
GreatHanGPT 古汉语数据集
数据集描述
这是一个用于训练古汉语大语言模型的数据集，包含从先秦到清初（1644-1722）的汉语文献。
数据来源
来源
内容
链接
chinese-poetry
唐诗宋词、楚辞、诗经、四书五经
GitHub
Werneror/Poetry
先秦到清末诗词，按朝代分
GitHub
CBETA
大正藏佛经
GitHub
数据规模
指标
数值
总记录数
2,400,939
总字符数
450,496,972
估计token数
~300M
时代分布
时代
记录数
字符数
占比
先秦
1,376
14,493,846
3.2%
汉魏
16,450
14,348,817
3.2%
隋唐
729,162
122,265,102
27.1%
两宋
728,569
97,907,260
21.7%… See the full description on the dataset page: https://huggingface.co/datasets/kenpusney/greathangpt-classical-chinese.taskweft-fbd-react-train
taskweft-fbd-react-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the compiler's reference scan on three constructed input traces per row. Every row is
constructed from a template and a seed, so the labels are true by… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-react-train.chinese-legal-sft
Chinese Legal SFT Dataset(中文法律 SFT 数据集)
面向大模型监督微调(SFT)的中文法律问答数据集,共 19,332 条问答对,
每条附带 LLM 质量评分。覆盖数据采集 → 清洗 → 去重 → 质量过滤 → 格式化 → 质量打分的完整数据工程流程。
配套代码与完整流水线:https://github.com/noah-white-python/legal-sft-dataset
数据构建流程
冷启动:基于开源数据集 DISC-Law-SFT 整理。
清洗:NFKC 全角半角统一、去控制字符、去空白、缺失过滤。
去重:精确去重(MD5)+ MinHash + LSH 近似去重(阈值 0.8)。
质量过滤:长度、中文字符占比等启发式规则,有效率 96.7%(20,000 → 19,332)。
格式化:输出标准 Alpaca 指令格式。
质量打分:用 LLM-as-judge 对全部数据从复杂度、清晰度、信息量三维度打分(1-5 分)。
字段说明… See the full description on the dataset page: https://huggingface.co/datasets/noah248/chinese-legal-sft.IndustryInstruction-Chinese
中文行业指令数据集
💻 Github Repo
简介
本数据集提取了原数据集 BAAI/IndustryInstruction 中源语言为中文的部分，并做了清洗。数据集分为单轮对话和多轮对话两个子集。
本数据集包含的行业及具体数据如下：
领域
单轮对话数目
多轮对话数目
AeroSpace
72667
0
Artificial-Intelligence
43906
0
Automobiles
78036
0
Finance-Economics
40135
0
Health-Medicine
177152
105320
Hospitality-Catering
39261
0
Law-Justice
43485
0
Literature-Emotions
44841
0
Subject-Education
271402
73
Technology-Research
41751
0
Transportation
51505
0
Travel-Geography
37150… See the full description on the dataset page: https://huggingface.co/datasets/Mxode/IndustryInstruction-Chinese.taskweft-fbd-editscore-train
taskweft-fbd-editscore-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the compiler and a runner that performed the plan. Every row is
constructed from a template and a seed, so the labels are true by construction and
the… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-editscore-train.taskweft-fbd-godot-train
taskweft-fbd-godot-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the engine itself: api_runner.gd performed the calls on the fixture scene and the returns were read back. Every row is
constructed from a template and a… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-godot-train.taskweft-fbd-trainer-train
taskweft-fbd-trainer-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the trainer config: the calls were applied to the mjlab task config and the term table read back. Every row is
constructed from a template and a seed… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-trainer-train.Chinese-DeepSeek-R1-Distill-data-110k-decontaminated
Decontaminated — Congliu/Chinese-DeepSeek-R1-Distill-data-110k
What this is
A filtered version of Congliu/Chinese-DeepSeek-R1-Distill-data-110k (revision
8520b649430617c2be4490f424d251d09d835ed3) with exact-duplicate rows and rows overlapping standard benchmark test sets
removed. This is a different artifact from the companion contamination report — that one is an
audit of what's wrong; this one is the corpus with those rows actually taken out, ready to train on.… See the full description on the dataset page: https://huggingface.co/datasets/liodon-ai/Chinese-DeepSeek-R1-Distill-data-110k-decontaminated.Chinese-Qwen3-235B-Thinking-2507-Distill-100k
📌 Note: The English translation of this dataset card is provided below.
Chinese-Qwen3-235B-Thinking-2507-Distill-100k
Dataset Summary
Chinese-Qwen3-235B-Thinking-2507-Distill-100k 是一个包含约 100k 条高质量中文推理与指令数据的数据集，由 Qwen-3-235B-A22B-Thinking-2507（官方 Thinking 模式，上下文长度 32K）蒸馏生成。
该数据集覆盖了多个重要领域：
数学与工程任务（Mathematics, Applied Math, Advanced Math）
通用知识与写作（General Knowledge, Language & Writing）
技术与编程（Technology & Programming）
商业与经济（Business & Economics）… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/Chinese-Qwen3-235B-Thinking-2507-Distill-100k.China-Halal-Restaurant
China Halal Restaurant Dataset (RAG Optimized) 🕌
This is a rigorously formatted Chinese Halal Restaurant corpus containing 201 authentic articles and travel guides. It is explicitly optimized for Retrieval-Augmented Generation (RAG) and pure text indexing. The data was explicitly designed to pass Hugging Face's Dataset Viewer standards natively by using optimal Parquet partitioning.
[!TIP]
Human Readers: Looking for the full text with all images perfectly rendered? Navigate to… See the full description on the dataset page: https://huggingface.co/datasets/qurancn/China-Halal-Restaurant.Chinese-Muslim-Travel
☪ Chinese-Muslim-Travel: Native Chinese Muslim Travel RAG Corpus
[!TIP]
Human Readers: Looking for the full text with all images perfectly rendered? Navigate to the Files and versions -> content folder to browse all Markdown articles natively!
Dataset Description
Chinese-Muslim-Travel is a curated RAG corpus containing 347 native Chinese articles documenting Muslim travel, halal food, mosque architecture, and Muslim community life across 20+ countries. Every… See the full description on the dataset page: https://huggingface.co/datasets/qurancn/Chinese-Muslim-Travel.taskweft-fbd-harness-train
taskweft-fbd-harness-train
Intents and the IEC 61131-3 Function Block Diagrams that carry them out, as an
EditScore-shaped corpus: one root row per intent, three candidates per row (rank1 the
reference diagram, rank3 one that compiles and does the wrong thing, rank5 one the
compiler refuses), and one score row per candidate from the compiler's reference scan, rank1's outputs the reference for the others. Every row is
constructed from a template and a seed, so the labels are true… See the full description on the dataset page: https://huggingface.co/datasets/chibifire/taskweft-fbd-harness-train.