datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
KodCode-Light-RL-10K
🐱 KodCode: A Diverse, Challenging, and Verifiable Synthetic Dataset for Coding
KodCode is the largest fully-synthetic open-source dataset providing verifiable solutions and tests for coding tasks. It contains 12 distinct subsets spanning various domains (from algorithmic to package-specific knowledge) and difficulty levels (from basic coding exercises to interview and competitive programming challenges). KodCode is designed for both supervised fine-tuning (SFT) and RL tuning.
🕸️… See the full description on the dataset page: https://huggingface.co/datasets/KodCode/KodCode-Light-RL-10K.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.conseil-detat-full-documents
Décisions du Conseil d'État (France)
Description
Ce dataset contient un corpus de décisions rendues par le Conseil d'État français, la plus haute juridiction de l'ordre administratif. Les décisions proviennent de la plateforme officielle Open Data de la Justice Administrative et sont diffusées au format XML anonymisé.
Le corpus rassemble les textes intégraux des décisions ainsi que plusieurs métadonnées permettant leur identification et leur traçabilité.
Source… See the full description on the dataset page: https://huggingface.co/datasets/hulk10/conseil-detat-full-documents.korean-bar-exam-hard-current-law-precedent-sft-1000
Korean Current-Law Bar Exam Hard SFT 1000
대한민국 현행 법령을 기준으로 만든 변호사시험 선택형 고난도 스타일 SFT 데이터 1,000문항입니다.
초기 직접 조문확인형 생성본은 실제 제14ㆍ15회 변호사시험보다 쉬워서, 이 버전은 다음 기준으로 다시 만들었습니다.
ㄱ/ㄴ/ㄷ/ㄹ 복합정오형 중심
甲/乙/丙, 검사ㆍ사법경찰관ㆍ행정청ㆍ회사ㆍ소송당사자 등이 등장하는 사례형 비중 확대
단순 근거 조문 선택형 제거
정답뿐 아니라 각 지문별 O/X 이유와 참고 법령 조문 제공
제15회 변호사시험 data/questions.csv와 높은 유사도 문항 제외
Files
data/questions.csv: Hugging Face preview용 메인 CSV입니다.
sft/train.jsonl: messages 형식 SFT용 JSONL입니다.
metadata/qa_report.json: 생성 수량, 난도 관련… See the full description on the dataset page: https://huggingface.co/datasets/gyung/korean-bar-exam-hard-current-law-precedent-sft-1000.RAIL-HH-10K RAIL-HH-10K: Multi-Dimensional Safety Alignment Dataset
The first large-scale safety dataset with 99.5% multi-dimensional annotation coverage across 8 ethical dimensions.
📖 Read Blog •
📖 Paper (Coming Soon) •
🚀 Quick Start •
🔌 RAIL API •
💻 Examples
🌟 What Makes RAIL-HH-10K Special?
🎯 Near-Complete Coverage
99.5% dimension coverage across all 8 ethical dimensions
Most existing datasets: 40-70% coverage
RAIL-HH-10K: 98-100%… See the full description on the dataset page: https://huggingface.co/datasets/responsible-ai-labs/RAIL-HH-10K.mmlu-auxiliary-train-10-choices
Dataset Card for Augmented MMLU (STEM) with Additional Distractors
Dataset Description
This dataset is an augmented version of the STEM portion of the MMLU auxiliary training set kz919/mmlu-auxiliary-train-auto-labelled, where each original 4-option multiple-choice question has been expanded to include 10 options (A-J) through the addition of six carefully constructed distractors.
Dataset Summary
Original Dataset: MMLU auxiliary training set (STEM portion)… See the full description on the dataset page: https://huggingface.co/datasets/andresnowak/mmlu-auxiliary-train-10-choices.python-text-copilot-training-instruct-ai-research-2024-02-10
Python Copilot Instructions on How to Code using Alpaca and Yaml
Training and test datasets for building coding multimodal models that understand how to use the open source GitHub projects for the multimodal Qwen AI project:
Qwen
Qwen Agent
Qwen VL Chat
Qwen Audio
This dataset is the 2024-02-10 update for the matlok python copilot datasets. Please refer to the Multimodal Python Copilot Training Overview for more details on how to use this dataset.
Details
Each row… See the full description on the dataset page: https://huggingface.co/datasets/matlok/python-text-copilot-training-instruct-ai-research-2024-02-10.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.ruler-100-nemotron
RULER-100 — Nemotron-Nano-v3 tokenized
RULER long-context evaluation data, regenerated with the
nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (instruct) tokenizer so the labeled context
lengths are exact for that model — instead of drifting, as they do when RULER data tokenized for a
different model (e.g. Qwen3) is fed to Nemotron.
What's here
7 context lengths: 4096, 8192, 16384, 32768, 65536, 131072, 262144 (the model's max).
13 RULER tasks: niah_single_1/2/3… See the full description on the dataset page: https://huggingface.co/datasets/jet-ai/ruler-100-nemotron.EarthScience-Text-LLM-20K-90-10
EarthScience-Text-LLM-20K-90-10
This is a pure-text Earth-science corpus unified from three non-overlapping upstream datasets:
Ekimetrics/climateqa-ipcc-ipbes-reports-1.0: climate and IPCC/IPBES report chunks.
GeoGPT-Research-Project/GeoGPT-CoT-QA: geoscience question-answer reasoning.
gremlin97/RemoteSensingCorpus: remote-sensing and geospatial machine-learning text.
Files and Split
The previous preprocessing outputs were merged into a 23,098-record pool and… See the full description on the dataset page: https://huggingface.co/datasets/moTcream/EarthScience-Text-LLM-20K-90-10.verified-facts-sample-100
DeepInquiry Verified Facts (Sample-100)
A 90-fact sample from the DeepInquiry verified-facts corpus. Every fact in this sample has been cross-checked against multiple structurally independent web sources, cited, dated, and confidence-scored before it entered the corpus.
This is a preview sample. The full corpus (~942 approved facts as of Sept 2026, growing continuously) is available via the DeepInquiry API at deepinquiry.ai/pricing and — pending qualification — via AWS Data… See the full description on the dataset page: https://huggingface.co/datasets/deepinquiry/verified-facts-sample-100.ds1000_pnyx
PNYX - DS-1000
This is a splitted and tested version of DS-1000, based on the reformatted version claudios/ds1000 (extracted metadata as columns). This version is designed to be compatible with the hf_evaluate code_eval package. Also, the code was modified to work with newer versions of the used python packages (numpy, scipy, etc.).
This dataset includes all the original fields and the following ones:
user_chat_prompt: A chat-style prompt for the problem, adapted from the prompt… See the full description on the dataset page: https://huggingface.co/datasets/PNYX/ds1000_pnyx.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… See the full description on the dataset page: https://huggingface.co/datasets/a13905873166/China-K12-STEM-10K-CoT-Reasoning.Liars_Are_Information
AIP — Adversarial Informativeness Pooling
"Liars Are Information": Byzantine-robust decentralized LLM swarms.
N agents answer the same question, broadcast an answer plus a confidence, and each
agent locally aggregates what it receives. A fraction f of agents are Byzantine.
Prior art (Krum, geometric median, trimmed mean, confidence weighting, Dawid–Skene,
filter-and-refine) discards adversarial input. AIP inverts coherent adversaries
and pools them as information — and then a… See the full description on the dataset page: https://huggingface.co/datasets/Dhruv1000/Liars_Are_Information.ppo_ppl_thinkfinal_stage10
PPO Stage-10 Curriculum Dataset
本仓库提供基于 Atomheart-Father/ppo_pool_24000_ppl10_sys10_thinkfinal_toklen 的分阶段 PPO/DPO 训练数据。源数据来自 OpenR1 子集与 OT-114k 数学子集，保持 <think>...</think><final>...</final> 的答案格式，并用 query-only PPL 做难度分桶。
数据切分
stage0 … stage9：共 10 个训练阶段。每阶段目标 2000 条（脚本参数 STAGE_SIZE=2000，FIXED_RATIO=0.8），约 80% 来自同难度分桶（stage_role=fixed），20% 为其他难度的混合样本（stage_role=random）。
eval：从剩余样本中采样（脚本参数 EVAL_SIZE=500），stage_role=eval。
test：剩余部分，stage_role=test。
stage_id：0–9 对应阶段，-1 表示… See the full description on the dataset page: https://huggingface.co/datasets/Atomheart-Father/ppo_ppl_thinkfinal_stage10.cgrt-consensus-5model
CGRT Consensus 5-Model Dataset
Multi-model consensus dataset for studying model agreement and disagreement patterns on mathematical reasoning tasks.
Dataset Description
61,678 math problems evaluated by 5 frontier LLMs with full reasoning traces and extracted answers.
Models Used
Model
Provider
Version
Claude
Anthropic
claude-3-5-sonnet-20241022
Codex/GPT-4
OpenAI
gpt-4o
Gemini
Google
gemini-1.5-flash
DeepSeek
DeepSeek
deepseek-chat
Qwen… See the full description on the dataset page: https://huggingface.co/datasets/Adam1010/cgrt-consensus-5model.ru-instruct-10k
10k Russian chatbot dialogues dataset
NCERT_Science_10thru-wikipedia-100k-full-text-daily-stats-10-years
📚 Russian Wikipedia Top 100K: Full Text with Daily Pageviews
**Крупнейший открытый датасет русскоязычной Википедии с полными текстами статей и ежедневной статистикой просмотров за 10 лет. МГУ, ОТиПЛ, 2025**
📖 Описание
Этот датасет содержит 99,348 самых популярных статей русскоязычной Википедии, отобранных по совокупному количеству просмотров за последние 10 лет.
Для каждой статьи собраны полный текст с сохранением структуры, метаданные и детальная ежедневная… See the full description on the dataset page: https://huggingface.co/datasets/Mikimi/ru-wikipedia-100k-full-text-daily-stats-10-years.magpie_llama70b_100k_swedish
Dataset Card
Llama 3.3 generated instruction: response pairs using the MagPie pipeline: https://github.com/magpie-align/magpie/ https://arxiv.org/abs/2406.08464
Larger and filtered dataset available here:
nicher92/magpie_llama70b_200k_filtered_swedish
Dataset Details
Dataset Description
System prompt template for generating instructions: "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nDu är en hjälpsam AI… See the full description on the dataset page: https://huggingface.co/datasets/nicher92/magpie_llama70b_100k_swedish.NCERT_Social_Studies_10th100MMLUpro
MMLU-Pro 100: A Balanced and Curated Evaluation Set
Dataset Description
This dataset is a curated, balanced subset of 100 questions derived from the TIGER-Lab/MMLU-Pro test set. It is designed to provide a small, fast, yet representative benchmark for evaluating the knowledge and reasoning capabilities of large language models across a wide range of academic and professional domains.
The key feature of this dataset is its stratified sampling method, ensuring that the… See the full description on the dataset page: https://huggingface.co/datasets/koiwave/100MMLUpro.felsefe-benchmark-100
Felsefe Benchmark — 100 Soru (Türkçe)
Yalnızca felsefe konularına odaklanan, 13 alt kategoriye yayılmış 100
soruluk çoktan seçmeli (A-E) bir değerlendirme seti. odev5_benchmark'taki
genel Türkçe MMLU testinden farklı olarak, bu set
yoitsmeyusuf/felsefe-lora
adaptörünü taban modeliyle ve farklı model aileleriyle karşılaştırmak için
hazırlandı.
Kategoriler
Yüzyıl Felsefesi: 7 soru
Yüzyıl Felsefesi: 1 soru
Antik Yunan Felsefesi: 11 soru
Epistemoloji: 9 soru… See the full description on the dataset page: https://huggingface.co/datasets/yoitsmeyusuf/felsefe-benchmark-100.birthday_quotes_1_to_100
Birthday Quote 1 to 100 — Full Combination Dataset
The Birthday Quote 1 to 100 dataset is an extensive collection of 3,807 birthday messages generated through complete combinations of tone, theme, and valid recipient types across realistic age groups.This dataset spans ages from 1 to 100 years, providing a highly diverse and customizable resource for generating personalized birthday wishes for any recipient.
Each entry contains a birthday message along with structured metadata — age… See the full description on the dataset page: https://huggingface.co/datasets/tejasashinde/birthday_quotes_1_to_100.korean-current-law-bar-exam-sft-1000
Korean Current-Law Bar Exam SFT 1000
대한민국 현행 법령을 기준으로 만든 변호사시험 선택형 스타일 SFT 데이터 1,000문항입니다.
이 데이터셋은 법무부 기출문제를 복제하지 않습니다. 기존 gyung/korean-bar-exam-moj-multiple-choice의 data/questions.csv는 난도와 과목 분포 참고 및 제15회 중복 방지 기준으로만 사용했습니다.
Files
data/questions.csv: Hugging Face preview용 메인 CSV입니다.
sft/train.jsonl: messages 형식 SFT용 JSONL입니다.
metadata/qa_report.json: 생성 수량, 과목 분포, 제15회 유사도 QA 결과입니다.
Columns
question_text: 문제와 5개 선택지
answer: 정답 번호, 1부터 5… See the full description on the dataset page: https://huggingface.co/datasets/gyung/korean-current-law-bar-exam-sft-1000.cleand_flatlander1024_or_instruct_dedup元データ: https://huggingface.co/datasets/flatlander1024/or_instruct_dedup
使用したコード: https://github.com/LLMTeamAkiyama/0-data_prepare/tree/master/src/flatlander1024-or_instruct_dedup
データ件数: 2,600
平均トークン数: 1,340
最大トークン数: 3,086
合計トークン数: 3,484,377
ファイル形式: JSONL
ファイル分割数: 1
合計ファイルサイズ: 13.0 MB
加工内容：
データセットの初期設定と読み込み:
flatlander1024/or_instruct_dedup データセットを読み込み、Pandas DataFrameに変換しました。
answer 列のデータ型を文字列 (str) に変換しました。
NLTKのpunktとstopwordsデータをダウンロードしました（必要な場合）。
IDの付与:… See the full description on the dataset page: https://huggingface.co/datasets/LLMTeamAkiyama/cleand_flatlander1024_or_instruct_dedup.covidqa-unique-context
Dataset Card for "covidqa-unique-context"
More Information needed
mobiplant
Dataset Card for MoBiPlant
Dataset Summary
MoBiPlant is a multiple-choice question-answering dataset curated by plant molecular biologists worldwide. It comprises two merged versions:
Expert MoBiPlant: 565 expert-level questions authored by leading researchers.
Synthetic MoBiPlant: 1,075 questions generated by large language models from papers in top plant science journals.
Each example consists of a question about plant molecular biology, a set of answer options, and… See the full description on the dataset page: https://huggingface.co/datasets/anonymousSub10/mobiplant.NCERT_Science_10thChemRD-Hard-100
ChemRD-Hard-100
100 bilingual (English / Chinese) PhD-level chemistry items, selected by measured difficulty from a
474-item verified pool. Every item is self-contained: everything needed to answer it is in the item.
Leaderboard
18 arms, each one isolated process or request per item, both languages, no shared context between
items. Reasoning is off wherever the endpoint allows it, and the last column reports the reasoning
tokens actually observed, so the claim is… See the full description on the dataset page: https://huggingface.co/datasets/EscheWang/ChemRD-Hard-100.