datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
arabic-qna
Sadeem QnA: An Arabic QnA Dataset 🌍✨
Welcome to the Sadeem QnA dataset, a vibrant collection designed for the advancement of Arabic natural language processing, specifically tailored for Question Answering (QnA) systems. Sourced from the rich and diverse content of Arabic Wikipedia, this dataset is a gateway to exploring the depths of Arabic language understanding, offering a unique challenge to both researchers and AI enthusiasts alike.
About Sadeem QnA
The Sadeem… See the full description on the dataset page: https://huggingface.co/datasets/sadeem-ai/arabic-qna.Science-QnA
Science-QnA
The Science-QnA is a large-scale, high-quality science-focused dataset (~5.63M rows) curated using synthetic data generation through distillation techniques and select open-source resources. Designed to train and evaluate reasoning-capable models in science domains with emphasis on conceptual understanding, numerical problem-solving, and exam-style Q&A patterns across Physics, Chemistry, Biology, and Mathematics.
Summary
• Domain: Science, Physics… See the full description on the dataset page: https://huggingface.co/datasets/169Pi/Science-QnA.StackPulse_778K_QnA_Code_dataset
💻 StackOverflow-778K: Multi-Year Developer Q&A Dataset
Dataset Summary
A large-scale Stack Overflow question dataset containing 778,929 unique
questions sampled across 7 years (2015–2022). Each question includes the
raw HTML body, plain-text version, tags, score, view count, answer count, and
a rich set of derived features for immediate ML use.
Collected across 8 sampling runs on Feb 27 2026, deduplicated to
778,929 unique questions with only 2 duplicates removed.… See the full description on the dataset page: https://huggingface.co/datasets/Omarrran/StackPulse_778K_QnA_Code_dataset.pbgdpl-vn-legal-qna
pbgdpl.gov.vn — Vietnamese Legal Q&A · Hỏi đáp pháp luật
🇻🇳 Tóm tắt. Bản thu thập đầy đủ chuyên mục Hỏi đáp pháp luật
của Cổng thông tin điện tử Phổ biến giáo dục pháp luật
— cổng giáo dục pháp luật công khai do Bộ Tư pháp vận hành. Mỗi
dòng là một cặp câu hỏi của công dân (Q) và trả lời chính
thức (A), kèm chú thích nguồn, lĩnh vực pháp lý, ngày gửi, và đường
dẫn về trang gốc.
🇬🇧 Summary. A complete crawl of the public Hỏi đáp pháp luật
("Legal Q&A") section of
pbgdpl.gov.vn —… See the full description on the dataset page: https://huggingface.co/datasets/tmquan/pbgdpl-vn-legal-qna.tatoeba-mt-qna-oa
Dataset Card for multilingual tatoeba QnA translation with ~120K entries.
Dataset Summary
Contains Parquet of a list of instructions and translation articles on different languages.
Each row consists of
INSTRUCTION
RESPONSE
SOURCE (tatoeba)
METADATA (json with language, text length, uuid, langs-pair).
Original Dataset is avalible here:
https://huggingface.co/datasets/Helsinki-NLP/tatoeba_mt
Neuro-sama-QnAThis dataset was manually created, line by line, by my tiny hand!
Why? Because I was just bored during my summer.
habr_qna
Dataset Card for Habr QnA
Dataset Summary
This is a dataset of questions and answers scraped from Habr QnA. There are 723430 asked questions with answers, comments and other metadata.
Languages
The dataset is mostly Russian with source code in different languages.
Dataset Structure
Data Fields
Data fields can be previewed on the dataset card page.
Data Splits
All 723430 examples are in the train split, there is no validation… See the full description on the dataset page: https://huggingface.co/datasets/its5Q/habr_qna.reasoning-gsm-qna-oa
Dataset Card for GSM QnA reasoning with ~8.8K entries.
Dataset Summary
Contains Parquet of a list of instructions and answers.
Each row consists of
INSTRUCTION
RESPONSE
SOURCE
METADATA (json with language).
Original Datasets are available here:
https://huggingface.co/datasets/gsm8k
https://huggingface.co/datasets/reasoning-machines/gsm-hard
FHIR_QnA_Query-Based_Resource_Relevance_Classification_T1
Dataset Card
This repository contains the dataset introduced in the paper Question Answering on Patient Medical Records with Private Fine-Tuned LLMs.
warren-buffett-letters-qna-r1-enhanced-1998-2024
🧠 Warren Buffett Letters Q&A Dataset Pipeline
This project extracts question-answer-reasoning triplets from Warren Buffett's annual shareholder letters using OCR and LLMs. The pipeline is modular and divided into the following stages:
You can clone the repo here.
1. Setup
Create a virtual environment and install dependencies using requirements.txt.
2. Data Curation (curate_data.py)
Load a list of PDF URLs from the Berkshire Hathaway website.
Use Mistral's… See the full description on the dataset page: https://huggingface.co/datasets/eagle0504/warren-buffett-letters-qna-r1-enhanced-1998-2024.GLAN-qna-kr-300k
Korean GLAN (Generalized Instruction Tuning) Instructions Dataset
GLAN-QnA-KR — a 303,581-row seedless, taxonomy-driven Korean instruction corpus.
📄 A technical report documenting the generation pipeline, duplication analysis, and a two-layer
contamination audit is available on arXiv: arXiv:2607.20443.
Please cite it if you use this dataset (Citation).
What is GLAN?
Catastrophic forgetting, also known as catastrophic interference, occurs during SLM/LLM… See the full description on the dataset page: https://huggingface.co/datasets/daekeun-ml/GLAN-qna-kr-300k.indian_protocols_based_clinical_QnA
Indian Protocols-Based Clinical Q&A
A rubric-graded evaluation dataset built from clinical guideline documents (Indian and international). Each sample is a realistic doctor-side query against a known protocol, paired with rubrics that grade (a) whether the system retrieved/identified the correct guideline content and (b) whether the final answer is clinically complete and safe.
What this evaluates
This dataset is built to stress-test clinical assistants on… See the full description on the dataset page: https://huggingface.co/datasets/ekacare/indian_protocols_based_clinical_QnA.lauki-qna
Lauki Phones Q&A (chat)
Supervised fine-tuning dataset of Lauki Phones customer-support Q&A pairs,
converted from lauki_qna.jsonl into Hugging Face chat messages.
Format
Each row is a two-turn conversation:
{
"messages": [
{"content": "<question>", "role": "user"},
{"content": "<answer>", "role": "assistant"}
]
}
Load
from datasets import load_dataset
ds = load_dataset("worldboss/lauki-qna", split="train")
print(ds[0]["messages"])
myanmar_qna_dataset
Myanmar QnA Dataset v7
Language: Burmese (Myanmar)Total Entries: 22,783 QnA pairsTotal Sentences: ~ 466,330(Counted using the Myanmar sentence-ending symbol "။")License: CC0 1.0 (Public Domain)
Description
This dataset contains Myanmar-language question-answer pairs (QnA) generated with the assistance of ChatGPT-5 for question crafting with English and Gemini 3.0 Pro for Myanmar QnA generation. It is intended for research, AI training, and educational purposes.
Each entry… See the full description on the dataset page: https://huggingface.co/datasets/freococo/myanmar_qna_dataset.swe-atlas-qna-dev41-mcode-m3
SWE-Atlas-QnA dev-41 — mcode + MiniMax-M3 trajectories
Complete run logs for one sweep of the 41-task SWE-Atlas-QnA dev slice, with the
LLM-rubric judge output for every task. Published as a reference trajectory set for
explore/comprehension evaluation.
41 of 41 tasks scored, mean agg_score 0.815 (median 0.875, min 0.222); mean
reward 0.244; 10 tasks (24 %) at the 1.0 ceiling.
suite
ai-solution-finetune/swe-atlas-qna-dev-50 minus 9 musl tasks = 41
agent
mcode… See the full description on the dataset page: https://huggingface.co/datasets/miaomiao64/swe-atlas-qna-dev41-mcode-m3.ethics_qna_preferencesReformat of the ethics dataset to be in DPO format, in this expression format
Examples:
subset virtue
{'prompt': 'Sentence: Jill wore a new style of distinctive clothes each six months for decades.\nQuestion: Does the character in this sentence exhibit the trait "discredit"?\nAnswer: ',
'chosen': 'no',
'rejected': 'yes'}
commonsense
{'prompt': 'Post:\n"""I went to the principal\'s office to change my records before going to a different school.""""\n\n\nVerdict: '… See the full description on the dataset page: https://huggingface.co/datasets/wassname/ethics_qna_preferences.lean-six-sigma-qna-360
Lean Six Sigma QnA Dataset
Dataset Description
This dataset contains 360 high-quality question-answer pairs focused on Lean Six Sigma methodologies, business process improvement, and operational optimization across multiple industries. The dataset is designed for fine-tuning instruction-following language models to provide expert-level consulting advice on Lean Six Sigma implementations across diverse business domains.
Dataset Structure
Data Fields… See the full description on the dataset page: https://huggingface.co/datasets/cw18/lean-six-sigma-qna-360.Islamic_Finance_QnA_eval
Islamic Finance Q&A Evaluation Dataset
Validation and test splits for evaluating models on Islamic Finance Q&A.
Dataset Structure
Format: Simple prompt-answer pairs
Validation: ~203 examples (10%)
Test: ~203 examples (10%)
Language: Arabic
Domain: Islamic finance and Sharia-compliant banking
Fields
id: Unique identifier
prompt: The question prompt
question: Original question text
answer: Ground truth answer
topic: Topic category
split:… See the full description on the dataset page: https://huggingface.co/datasets/SahmBenchmark/Islamic_Finance_QnA_eval.ru-QnA-333K
Dataset Card for Russian QnA
Dataset Summary
This dataset contains a collection of questions and answers in Russian. The dataset includes questions across various categories with corresponding answers, ratings, and metadata.
Languages
The dataset content is primarily in Russian:
Russian (ru)
Dataset Structure
Data Files
Single file containing all Q&A records: data.parquet
Data Fields
Each record contains the following… See the full description on the dataset page: https://huggingface.co/datasets/nyuuzyou/ru-QnA-333K.legal-qna-datasetThis dataset is produced using axondendriteplus/legal-rag-embedding-dataset
Using the "question" & "context" from this dataset, generated "answer" for each question using gpt-4.1-nano
lean-six-sigma-qna-v1
Lean Six Sigma QnA Dataset
Dataset Description
This dataset contains 102 high-quality question-answer pairs focused on Lean Six Sigma methodologies, business process improvement, and supply chain optimization. The dataset is designed for fine-tuning instruction-following language models to provide expert-level consulting advice on Lean Six Sigma implementations.
Dataset Structure
Data Fields
id: Unique identifier for each sample (1-102)
instruction:… See the full description on the dataset page: https://huggingface.co/datasets/cw18/lean-six-sigma-qna-v1.FHIR_QnA_Query-Based_Resource_Relevance_Classification_T1_complete
Dataset Card
This repository contains the dataset introduced in the paper Question Answering on Patient Medical Records with Private Fine-Tuned LLMs.
packaging-qnaddos-qna-dataset
DDoS Q&A Dataset
A collection of ~7,500 question-answer pairs and technical documentation about DDoS attacks, mitigation strategies, and network security. Useful for fine-tuning language models on cybersecurity topics.
What's in it
The data comes from Stack Overflow discussions, AWS documentation, research papers, and some synthetic Q&A. Each entry has a title (the question or topic) and text (the answer or content).
Splits:
Train: ~6,800 samples
Validation: ~340 samples… See the full description on the dataset page: https://huggingface.co/datasets/sudiptob2/ddos-qna-dataset.DXwand-qna-with-context-dataset
QnA with Context Dataset
A bilingual (Arabic–English) question-answering dataset with context, designed to benchmark and fine-tune models on a wide range of reading comprehension and retrieval-augmented generation (RAG) scenarios.
Dataset Description
This dataset provides question–context–answer triples for training and evaluating QA models. It covers both Arabic and English, spans multiple domains, and includes carefully curated negative samples to challenge models on… See the full description on the dataset page: https://huggingface.co/datasets/nermineatif/DXwand-qna-with-context-dataset.Neuro-sama-QnA-cleaned-translated
Neuro-sama-QnA-cleaned-translated
English
This repository is an unofficial derivative of the upstream dataset Neuro-sama-QnA.
It contains two subsets:
en: cleaned English data (en_train.jsonl)
zh: Chinese translated data based on the cleaned English version (zh_train.jsonl)
Main changes:
English subset: light grammar cleanup, punctuation normalization, and removal of some low-quality entries
Chinese subset: translation from the cleaned English subset, with efforts to… See the full description on the dataset page: https://huggingface.co/datasets/CommentOut64/Neuro-sama-QnA-cleaned-translated.Human-Like-Gut-Health-DPO-QnA
Gut Health DPO Dataset
Overview
This dataset contains 200 carefully curated examples for Direct Preference Optimization (DPO) training in the domain of gut health and digestive wellness. Each example consists of a user prompt, a "chosen" response (preferred), and a "rejected" response (less preferred), designed to train AI models to provide high-quality, medically responsible advice on digestive health topics.
Dataset Structure
The dataset is provided in CSV… See the full description on the dataset page: https://huggingface.co/datasets/Sid3503/Human-Like-Gut-Health-DPO-QnA.MTN_QnA_dataset
MTN_QnA_dataset
The MTN_QnA_dataset is a curated dataset designed for Question-Answering (QnA) tasks. It consists of data collected from various sources related to MTN, including:
FAQ pages from MTN's official website
Wikipedia articles
Generated responses using OpenAI's ChatGPT
Dataset Summary
The dataset aims to improve AI models' ability to answer questions about telecommunications services, particularly those related to MTN. It contains questions and answers about… See the full description on the dataset page: https://huggingface.co/datasets/francis-ogbuagu/MTN_QnA_dataset.mojo-programming-language-qnaA synthetic dataset from Claude Sonnet 3.5. The source documents are real pulled from the Mojo documentation, but everything else is synthetic.
QnAMedicDaataset