datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
sql-create-context
Overview
This dataset builds from WikiSQL and Spider.
There are 78,577 examples of natural language queries, SQL CREATE TABLE statements, and SQL Query answering the question using the CREATE statement as context. This dataset was built with text-to-sql LLMs in mind, intending to prevent hallucination of column and table names often seen when trained on text-to-sql datasets. The CREATE TABLE statement can often be copy and pasted from different DBMS and provides table names, column… See the full description on the dataset page: https://huggingface.co/datasets/b-mc2/sql-create-context.context_qa_sum_qwen3_synthetic
Context-based QA and Summarization Synthetic Dataset
Overview
This dataset contains synthetic context-based question-answering (QA) and summarization data. The data was synthesized using:
Source context: openbmb/Ultra-FineWeb
Synthesis model: Qwen3-30B-A3B-Instruct-2507
Each context is obtained by taking the initial segment of raw pretraining text from Ultra-FineWeb, truncated to at most the corresponding number of tokens, while ensuring the truncation does not occur in… See the full description on the dataset page: https://huggingface.co/datasets/yuyijiong/context_qa_sum_qwen3_synthetic.ledger-long-context-KPI-QA
LEDGER — Long-Context KPI Question Answering & Page Retrieval
This dataset is part of the LEDGER (Long-context Evaluation of Documents for
Grounded Extraction and Retrieval) benchmark.
It supports two of the three LEDGER tasks:
Page-level KPI retrieval — given a natural-language question about a financial
KPI and the corresponding annual report, retrieve the relevant page(s). Each row
includes TREC-style graded relevance judgments (qrels) over all candidate pages.… See the full description on the dataset page: https://huggingface.co/datasets/artefactory/ledger-long-context-KPI-QA.long-context-qa-curated-20
Dataset Card / 数据集卡
Dataset Description / 数据集简介
This public release contains 20 curated samples selected from a 10,000-record long-context QA collection. It targets retrieval over long documents, cross-section evidence synthesis, numerical reasoning, timeline reconstruction, and structured answer evaluation. The public subset contains 15 short-answer questions and 5 multiple-choice questions, balanced across Chinese and English.
本公开版本从 10,000 条长上下文问答数据中精选 20… See the full description on the dataset page: https://huggingface.co/datasets/LianeMarilin/long-context-qa-curated-20.Agentic-Long-Context-Understanding-QA 📖 Agentic Long Context Understanding 📖
Self-Taught Agentic Long Context Understanding (Arxiv).
AgenticLU refines complex, long-context queries through self-clarifications and contextual grounding, enabling robust long-document understanding in a single pass.
Installation Requirements
This codebase is largely based on OpenRLHF and Helmet, kudos to them.
The requirements are the same
pip install openrlhf
pip install -r ./HELMET/requirements.txt… See the full description on the dataset page: https://huggingface.co/datasets/yzhuang/Agentic-Long-Context-Understanding-QA.context
On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
Charlie Zhang, Graham Neubig,
Xiang Yue
Carnegie Mellon University, Language Technologies Institute
Does Reinforcement Learning Truly Extend Reasoning?
This work explores the discrepancy in views on RL's effectiveness in extending language models' reasoning abilities. Some characterize RL as a capability refiner, while others see it as inducing new compositional skills. This challenge… See the full description on the dataset page: https://huggingface.co/datasets/Interplay-LM-Reasoning/context.Multi-turn_Long-context_Benchmark_for_LLMs
LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues
Arxiv: https://www.arxiv.org/abs/2507.13681
Huggingface: https://huggingface.co/papers/2507.13681
Introduction
LoopServe Multi-Turn Dialogue Benchmark is a comprehensive evaluation dataset comprising multiple diverse datasets designed to assess large language model performance in realistic conversational scenarios.
Unlike traditional benchmarks that place queries only at the end… See the full description on the dataset page: https://huggingface.co/datasets/TreeAILab/Multi-turn_Long-context_Benchmark_for_LLMs.open-ended-w-context
Dataset Summary
EVE-open-ended-w-context is a collection of open-ended question-answer pairs focused on Earth Observation (EO) with accompanying context documents. Unlike the standard open-ended dataset, this version provides up to 3 relevant documents for each question that models can use to ground their responses. This makes it ideal for evaluating Retrieval-Augmented Generation (RAG) systems and testing models' ability to leverage provided context when answering questions.
The… See the full description on the dataset page: https://huggingface.co/datasets/eve-esa/open-ended-w-context.ascp-context-attribution
ASCP: Causal Context Attribution and Probe Benchmark
Released artifacts for The Laws of Context Allocation: Causal Measurement and
Closed-Loop Orchestration in Generative Search.
📄 Paper: https://arxiv.org/abs/2608.23252
💻 Code: https://github.com/PeiYangLiu/ascp
Retrieval-augmented generation is usually measured with relevance proxies —
BM25, query–document cosine, output overlap — that score how related a passage
looks, not whether the generator used it. This dataset ships… See the full description on the dataset page: https://huggingface.co/datasets/PeiyangLiu/ascp-context-attribution.quran-question-answer-context
Dataset Card for "quran-question-answer-context"
Dataset Summary
Translated the original dataset from Arabic to English and added the Surah ayahs to the context column.
Usage
from datasets import load_dataset
dataset = load_dataset("nazimali/quran-question-answer-context")
DatasetDict({
train: Dataset({
features: ['q_id', 'question', 'answer', 'q_word', 'q_topic', 'fine_class', 'class', 'ontology_concept', 'ontology_concept2', 'source', 'q_src_id'… See the full description on the dataset page: https://huggingface.co/datasets/nazimali/quran-question-answer-context.jaqket_v1_qa_wikija_context
⚠️ ! 注意 !
公開したデータセットは、私の不注意で本来学習させる用途には相応しくない、JAQKET テスト用のデータセットも train に含んでしまっています。
そのため、以下のように再度テスト用のデータを分離してからの利用をお勧めします。
import datasets
ds = datasets.load_dataset("hotchpotch/jaqket_v1_qa_wikija_context")
train_ds = ds["train"] # type: ignore
valid_ds = ds["validation"] # type: ignore
# 大元の "hotchpotch/jaqket_v1_qa_wikija_context" の train / valid 分割方法がよくないので、再度分割する
train_df = train_ds.to_pandas()
valid_df = valid_ds.to_pandas()
df = pd.concat([train_df, valid_df])… See the full description on the dataset page: https://huggingface.co/datasets/hotchpotch/jaqket_v1_qa_wikija_context.sql-create-context-pt
Overview
Este dataset é uma versão traduzida para o português do dataset b-mc2/sql-create-context,
que foi construído a partir dos datasets WikiSQL e Spider. Ele contém exemplos de perguntas
em português, instruções SQL CREATE TABLE e consultas SQL que respondem às perguntas
utilizando a instrução CREATE TABLE como contexto.
O principal objetivo deste dataset é ajudar modelos de linguagem natural em português a gerar consultas
SQL precisas e contextualizadas, prevenindo a… See the full description on the dataset page: https://huggingface.co/datasets/emdemor/sql-create-context-pt.context-primitive-code-agent-pack-v0
Context Primitive Code-Agent Pack v0 — Free Funnel
Free product-specific instruction / Q&A seed material from Primitive Origins’ Context Primitive / Foundry tests.
This is a marketing / companion corpus for the Context Primitive stack — not a general public code-agent marketplace hero SKU.
What’s inside
JSONL splits under data/:
behavior_qa.train.jsonl / .eval.jsonl
instruction_test_generation.train.jsonl / .eval.jsonl
foundry/python_test_generation.*… See the full description on the dataset page: https://huggingface.co/datasets/Primitive-Origins/context-primitive-code-agent-pack-v0.medhallu-twins-repaired-context
MedHallu twins with repaired context
Balanced medical hallucination-detection twins for training a small model to
detect hallucinated answers and explain why. Each row is a
(question, answer, context) triple labelled row_type.
Built from MedHallu, which pairs -- for the same question and source -- a
correct Ground Truth answer with a planted Hallucinated Answer. We keep
both as a twin pair, so within a pair the only difference is the hallucination.
That removes the… See the full description on the dataset page: https://huggingface.co/datasets/Certops/medhallu-twins-repaired-context.hausa-stem-reasoning-with-cultural-context
Hausa STEM Reasoning with Cultural Context
Abstract
We present the first large-scale bilingual Hausa-English STEM reasoning dataset with deep cultural adaptation, containing 2,640 high-quality question-answer pairs translated from the STEM-Reasoning-Complex dataset. Our work introduces the "Shehin Malamin Kimiyya" (The Wise Scholar of Science) translation framework, which transforms Western scientific concepts into culturally-embedded Hausa explanations using systematic… See the full description on the dataset page: https://huggingface.co/datasets/Tushe/hausa-stem-reasoning-with-cultural-context.sql-create-context-instruction
Overview
This dataset is built upon SQL Create Context, which in turn was constructed using data from WikiSQL and Spider.
There are 78,577 examples of natural language queries, SQL CREATE TABLE statements, and SQL Query answering the question using the CREATE statement as context. This dataset was built with text-to-SQL LLMs in mind, intending to prevent hallucination of column and table names often seen when trained on text-to-SQL datasets. The CREATE TABLE statement can often be… See the full description on the dataset page: https://huggingface.co/datasets/bugdaryan/sql-create-context-instruction.sql-create-context-copy
Fork of b-mc2/sql-create-context
Overview
This dataset builds from WikiSQL and Spider.
There are 78,577 examples of natural language queries, SQL CREATE TABLE statements, and SQL Query answering the question using the CREATE statement as context. This dataset was built with text-to-sql LLMs in mind, intending to prevent hallucination of column and table names often seen when trained on text-to-sql datasets. The CREATE TABLE statement can often be copy and pasted from… See the full description on the dataset page: https://huggingface.co/datasets/philschmid/sql-create-context-copy.context-as-a-service
CaaS Benchmark Corpus v1
A diverse collection of synthetic enterprise documents for benchmarking context extraction and RAG systems.
Dataset Description
This dataset contains 16 representative enterprise documents spanning multiple formats and domains, designed to evaluate:
Structure-aware indexing - Can the system identify high-value vs. low-value content?
Time decay relevance - Does the system properly weight recent vs. old information?
Pragmatic truth detection - Can… See the full description on the dataset page: https://huggingface.co/datasets/imran-siddique/context-as-a-service.Financial_Context_DatasetThis dataset contains over 50,000 samples of user financial queries paired with their corresponding structured data requests (context). It was created to facilitate the creation of the Financial Agent LLM for accurate data extraction and query answering.
How to load the Dataset
You can load the dataset using the code below:
from datasets import load_dataset
ds = load_dataset("Chaitanya14/Financial_Context_Dataset")
Dataset Construction
Diverse Query Sources… See the full description on the dataset page: https://huggingface.co/datasets/Chaitanya14/Financial_Context_Dataset.pandas-create-context
Overview
This dataset is built from sql-create-context, which in itself builds from WikiSQL and Spider.
I have used GPT4 to translate the SQL schema into pandas DataFrame schem initialization statements and to translate the SQL queries into pandas queries.
There are 862 examples of natural language queries, pandas DataFrame creation statements, and pandas query answering the question using the DataFrame creation statement as context. This dataset was built with text-to-pandas… See the full description on the dataset page: https://huggingface.co/datasets/hiltch/pandas-create-context.sql-create-context-id
Overview
This dataset is a fork from sql-create-context
This dataset builds from WikiSQL and Spider.
There are 78,577 examples of natural language queries, SQL CREATE TABLE statements, and SQL Query answering the question using the CREATE statement as context. This dataset was built with text-to-sql LLMs in mind, intending to prevent hallucination of column and table names often seen when trained on text-to-sql datasets. The CREATE TABLE statement can often be copy and pasted from… See the full description on the dataset page: https://huggingface.co/datasets/detakarang/sql-create-context-id.truthful_qa_context
Dataset Card for truthful_qa_context
Dataset Summary
TruthfulQA Context is an extension of the TruthfulQA benchmark, specifically designed to enhance its utility for models that rely on Retrieval-Augmented Generation (RAG). This version includes the original questions and answers from TruthfulQA, along with the added context text directly associated with each question. This additional context aims to provide immediate reference material for models, making it particularly… See the full description on the dataset page: https://huggingface.co/datasets/portkey/truthful_qa_context.German-RAG-ORPO-Long-Context-ShareGPT-HESSIAN-AI
German-RAG-ORPO (Odds Ratio Preference Optimization) Long Context ShareGPT-Format
German-RAG - German Retrieval Augmented Generation
Dataset Summary
The ORPO Long Context Tasks Dataset represents a specialized collection for fine-tuning language models with a focus on RAG-specific capabilities.
The subsets are derived from Synthetic generation inspired by Tencent's (“Scaling Synthetic Data Creation with 1,000,000,000 Personas”).
Dataset Structure… See the full description on the dataset page: https://huggingface.co/datasets/avemio/German-RAG-ORPO-Long-Context-ShareGPT-HESSIAN-AI.ContextMATH
ContextMATH
ContextMATH is a benchmark dataset for evaluating the contextual mathematical reasoning of Large Language Models (LLMs). It is introduced in the paper:
From Abstract to Contextual: What LLMs Still Cannot Do in MathematicsAccepted at ICLR 2026📄 https://arxiv.org/abs/2601.23048
Dataset Description
While LLMs perform impressively on abstract math benchmarks, their accuracy drops significantly when the same problems are embedded in real-world narrative… See the full description on the dataset page: https://huggingface.co/datasets/bwcao/ContextMATH.German-RAG-ORPO-Long-Context-Alpaca-HESSIAN-AI
German-RAG-ORPO (Odds Ratio Preference Optimization) Long-Context Alpaca-Format
German-RAG - German Retrieval Augmented Generation
Dataset Summary
The ORPO Long Context Tasks Dataset represents a specialized collection for fine-tuning language models with a focus on RAG-specific capabilities.
The subsets are derived from Synthetic generation inspired by Tencent's (“Scaling Synthetic Data Creation with 1,000,000,000 Personas”).
Dataset Structure… See the full description on the dataset page: https://huggingface.co/datasets/avemio/German-RAG-ORPO-Long-Context-Alpaca-HESSIAN-AI.sql-create-context-thai
Overview
This dataset builds from sql-create-context.
@misc{b-mc2_2023_sql-create-context,
title = {sql-create-context Dataset},
author = {b-mc2},
year = {2023},
url = {https://huggingface.co/datasets/b-mc2/sql-create-context},
note = {This dataset was created by modifying data from the following sources: \cite{zhongSeq2SQL2017, yu2018spider}.},
}
Contextualized_Privacy_Defense_Trajectory
Contextualized Privacy Defense
Paper: Contextualized Privacy Defense for LLM Agents
Code: https://github.com/SALT-NLP/contextual_privacy_defense
Abstract:
Abstract LLM agents increasingly act on users’ personal information, yet existing privacy defenses remain limited in both design and adaptability. Most prior approaches rely on static or passive defenses, such as prompting and guarding. These paradigms are insufficient for supporting contextual, proactive privacy… See the full description on the dataset page: https://huggingface.co/datasets/SALT-NLP/Contextualized_Privacy_Defense_Trajectory.ContextAwareResponses
Context Aware Responses
This is a synthetically generated dataset using Llama 3.3 and Tavily search.
At the moment, the dataset is generated using the sentence-transformers/natural-questions dataset as a base, taking the queries from that dataset and feeding it into a context fetching and data generation pipeline.
Synthetic-Context-Conversations
Synthetic-Context-Conversations
Overview
The Synthetic-Context-Conversations dataset is a collection of synthetic conversations designed to simulate empathetic and context-rich dialogues. It is particularly useful for tasks such as text generation, summarization, and question answering. The dataset is available in English and contains between 10,000 to 100,000 entries.
Dataset Details
Modalities: Text
Languages: English
Size: 10K-100K
Formats: Parquet
License:… See the full description on the dataset page: https://huggingface.co/datasets/prithivMLmods/Synthetic-Context-Conversations.tinyllama-context-retrieval