datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
olympiad_style_integer_math_problems
Olympiad Math Corpus
Version: v2.1.1
Release date: 2026-05-03
59,486 synthetically generated olympiad-style math problems with verified integer answers and formal computation graphs.
Loading
from datasets import load_dataset
ds = load_dataset("mihailgribov/olympiad_style_integer_math_problems", split="train")
lemma_applicability is stored as list[{lemma, status}] rather than a sparse dict (required for Arrow-based consumers). To convert to a dict for local use:… See the full description on the dataset page: https://huggingface.co/datasets/mihailgribov/olympiad_style_integer_math_problems.olympiad_style_integer_math_reasoning
Olympiad Math Reasoning Traces
Version: v1.0.2
Release date: 2026-04-19
64,763 full model reasoning traces for olympiad-style math problems with verified integer answers. This dataset contains only correct and non-truncated traces — every record contains a terminal \boxed{...} answer (within the last 500 characters of the response) that matches the expected integer exactly, and none of the responses hit the model's generation-token cap. Intended for distillation and supervised… See the full description on the dataset page: https://huggingface.co/datasets/mihailgribov/olympiad_style_integer_math_reasoning.StyleCruxGen
Dataset Card for StyleCruxGen
A visual dataset exploring the interplay between style, object, and environment using diffusion-generated imagery.
Dataset Details
Dataset Description
StyleCruxGen is a synthetic multi-style image dataset generated using the Stable Diffusion XL (SDXL) model.
It contains 15465 high-resolution images (1024px) featuring 30 distinct object-environment pairs rendered across photorealistic and 10 artistic styles.
The 30… See the full description on the dataset page: https://huggingface.co/datasets/bodhisattamaiti/StyleCruxGen.Associated-Press-Stylebook
Federal Travel Regulation
Maintainer: Terry Eppler
Owner: US Federal Government
Dataset Summary
This dataset contains document-grounded question-and-answer records based on the Federal Travel Regulation, as reproduced in the source volume of title 41 of the Code of Federal Regulations.
The Federal Travel Regulation establishes government-wide policies governing official civilian travel and relocation at Federal expense. It addresses temporary duty travel… See the full description on the dataset page: https://huggingface.co/datasets/leeroy-jankins/Associated-Press-Stylebook.Alpha_Chat_Style_Dataset
🦾 Alpha Chat Style Dataset | darkknight25
Inject dominance, charm, and precision into your LLMs.
Crafted by Sunny Thakur, this dataset is designed to train conversational agents that speak like a leader, think like a tactician, and respond like a professional.
“Control the tone. Command the room. Every word should land like a calculated move.” – Alpha Protocol
🎯 Purpose
This dataset enables large language models—like Mixtral 8x7B Instruct—to adopt a bold… See the full description on the dataset page: https://huggingface.co/datasets/darkknight25/Alpha_Chat_Style_Dataset.japanese-style-contrast-dataset
LLMの文体差理解向上のための日本語データセット
本データセットは、新居浜工業高等専門学校と株式会社APTOの共同研究により構築された、日本語LLMの文体差（話し言葉・書き言葉）理解性能の向上を目的とした学習用データセットです。
新居浜高専 電子工学専攻の高専生が研究の企画立案からデータセット構築、追加学習および評価実験までを主導し、指導教員の助言と株式会社APTOからの計算環境の提供・技術的指導のもとで作成されました。
本研究の成果は言語処理学会第32回年次大会(NLP2026)にて発表されます。
研究の背景
大規模言語モデル（LLM）は日本語の質問応答や文章生成において高い性能を示す一方で、話し言葉と書き言葉といった文体差に対する理解には課題が残されております。特に話し言葉では、口語的表現や省略を含むため理解性能が不安定になりやすく、文体差への対応は日本語LLMにおける重要な課題となっています。… See the full description on the dataset page: https://huggingface.co/datasets/APTO-001/japanese-style-contrast-dataset.gpt3.5-jinyong-style
OpenAI微调医学数据集
开源数据集存在与预训练模型数据集相似度较高的问题，导致微调结果不明显，效果不突出，因此就需要构建某种风格的专属数据集。
因此，构建了带金庸创作风格的医学数据集，可直接用于openai的微调使用。
微调实验结果证明，当输入医学问题时，微调模型回答的风格为金庸创作风格回答。
Question
[{'role': 'system', 'content': '你是我的私人医生助手，你要回答我的健康问题。'}, {'role': 'user', 'content': '我睡眠不足，很头痛'}]
Answer
少侠保重身体。江湖人称“睡眠不足，头痛难受”，此乃常见之症。吾乃医术高强之侠士，愿为你指点迷津。睡眠不足，多因忧虑纷扰、劳累过度所致。当以调整生活作息，保持规律之睡眠时间为要。且可饮用一些花草茶，如菊花、薰衣草，以宁心安神。如症状持续，还需请教名医，以寻得更为妥善之疗法。
可以发现，模型的回答已经学习到金庸创作风格的能力。
korean-ai-prompt-style-dataset
Korean AI Prompt Style Dataset
같은 질문에 대해 ChatGPT, Gemini, Claude, Copilot 각각에 최적화된 프롬프트 스타일을 비교한 한국어 데이터셋입니다.
A Korean dataset comparing optimized prompt styles for ChatGPT, Gemini, Claude, and Copilot on identical questions.
📌 데이터 구조 / Data Structure
```json
{
"instruction": "질문 내용 / user question",
"input": "target: chatgpt / gemini / claude / copilot",
"output": "해당 AI에 최적화된 프롬프트 / optimized prompt for the target AI"
}
```
📊 데이터 현황 / Stats… See the full description on the dataset page: https://huggingface.co/datasets/podongchip/korean-ai-prompt-style-dataset.guanaco-sharegpt-style-serbian
Guanaco Sharegpt-style Serbian
Dataset Description
This dataset is a Serbian-translated version of the philschmid/guanaco-sharegpt-style
Dataset Structure
Usage
To load the dataset in Serbian, run:
from datasets import load_dataset
ds = load_dataset("datatab/guanaco-sharegpt-style-serbian")
Data Splits
The dataset has one splits, suitable for:
Supervised fine-tuning (sft).
The dataset is stored in parquet format with each entry using… See the full description on the dataset page: https://huggingface.co/datasets/datatab/guanaco-sharegpt-style-serbian.llama2-jinyong-style
LLama2微调医学数据集
开源数据集存在与预训练模型数据集相似度较高的问题，导致微调结果不明显，效果不突出，因此就需要构建某种风格的专属数据集。
因此，构建了带金庸创作风格的医学数据集，可直接用于LLama2的微调使用。
微调实验结果证明，当输入医学问题时，微调模型回答的风格为金庸创作风格回答。
Instruction
Instruction: 我睡眠不足，很头痛
Output
少侠保重身体。江湖人称我为医仙，今日得闻君有头痛之症，实非善兆。睡眠不足，乃是身体之不适，需慎之又慎。头痛之症，或许源自于血液循环不畅，或许源于神经压力过大。当务之急，当调整生活习惯，保持良好的睡眠规律，避免过度劳累。此外，可尝试调整饮食，避免辛辣刺激之物，以免加重头痛之苦。如君仍遭此病痛，可寻求名医良药，以求解忧。</s>
可以发现，模型的回答已经学习到金庸创作风格的能力。
persian-math-alpaca-style
Persian Math in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the Math original dataset which is translated to Persian in this repository.
Conversion-of-Fomality-and-Style
🇰🇿 Kazakh Formal Style Transfer and Register Adaptation
📖 Overview
This dataset contains 1,500 samples focused on the task of Style Transfer, specifically converting casual or descriptive Kazakh text into a professional, "Business/Official" register (Іскерлік стиль).
📊 Dataset Statistics
General Metrics
Metric
Count
Total Samples
1,500
Total Words (approx.)
134,949
Avg. Words per Sample
90
Word… See the full description on the dataset page: https://huggingface.co/datasets/farabi-lab/Conversion-of-Fomality-and-Style.odia-gemma4-style-polish-mix
OdiaEdgeVoice Gemma4 Style Polish Mix
Weighted dataset for improving Odia chat behavior, punctuation, concise answering,
Romanized Odia handling, and refusal behavior.
Reference runtime model: kaushikdash/odia-gemma4-e2b-gguf
Training base used by notebook: google/gemma-4-E2B-it
Important: GGUF artifacts are not directly trainable. This dataset is intended for
LoRA fine-tuning the trainable Gemma base, then exporting/quantizing back to GGUF.
Target Mix
{… See the full description on the dataset page: https://huggingface.co/datasets/kaushikdash/odia-gemma4-style-polish-mix.parsinlu-reading-comprehension-alpaca-style
ParsiNLU Reading Comprehension in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ParsiNLU original dataset.
parsinlu-entailment-alpaca-style
ParsiNLU Entailment in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ParsiNLU original dataset.
parsinlu-multiple-choice-alpaca-style
ParsiNLU Multiple Choice in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ParsiNLU original dataset.
StylExNet5k
Dataset Card for StylExNet5k
StylExNet5k is a multi-style synthetic image dataset consisting of 5,000 images across 100 everyday object categories, each rendered in 10 distinct artistic or representational styles and placed in varied real-world contextual environments.
It is intended to support evaluation and training of computer vision and vision–language models across style and context domains.
Dataset Details
Dataset Description
StylExNet5k is a synthetic… See the full description on the dataset page: https://huggingface.co/datasets/bodhisattamaiti/StylExNet5k.HILL_Learning-style_JailbreakPlease read hill_learning-style_jailbreak-dataset-license-agreement
For details, please refer to the paper: A Simple and Efficient Jailbreak Method Exploiting LLMs' Helpfulness
We reveal a critical safety blind spot in modern LLMs: learning-style queries, which closely resemble ordinary educational questions, can reliably elicit harmful responses.
Our HILL framework introduces a novel and systematic method for constructing such queries.
HILL achieves high attack success rates with great… See the full description on the dataset page: https://huggingface.co/datasets/gracehuggingface/HILL_Learning-style_Jailbreak.farstail-entailment-alpaca-style
FarsTail Entailment in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the FarsTail original dataset.
persian-conjnli-entailment-alpaca-style
Persian ConjNLI Entailment in Alpaca Style
This dataset is an Alpaca-style and instruction-included version of the ConjNLI which is translated to Persian in this repository.
Life-style