datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
SCBench
SCBench
[Paper]
[Code]
[Project Page]
SCBench (SharedContextBench) is a comprehensive benchmark to evaluate efficient long-context methods in a KV cache-centric perspective, analyzing their performance across the full KV cache lifecycle (generation, compression, retrieval, and loading) in real-world scenarios where context memory (KV cache) is shared and reused across multiple requests.
🎯 Quick Start
Load Data
You can download and load the SCBench data… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/SCBench.Taskbench
TaskBench: Benchmarking Large Language Models for Task Automation
Introduction
TaskBench is a benchmark for evaluating large language models (LLMs) on task automation. Task automation can be formulated into three critical stages: task decomposition, tool invocation, and parameter prediction. This complexity makes data collection and evaluation more challenging compared to common NLP tasks. To address this challenge, we propose a comprehensive evaluation framework… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/Taskbench.XL-DocBench
XL-DocBench
Evidence-grounded reasoning across hundreds or thousands of pages.
Fully verified by 194 human experts.
Hongchen Wei1,†,‡, Yuanzhe Wang2,†,‡,
Bei Liu2,*, Yifan Yang2, Qi Dai2,
Ruichun Ma2, Kai Qiu2, Yunsheng Li2,
Dongdong Chen2, Chong Luo2,
Zhenzhong Chen1, Baining Guo2
1Wuhan University 2Microsoft
†Equal contribution ‡Work done during an internship at MSRA
*Project leader
Project Page ·
Paper ·
Live Leaderboard… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/XL-DocBench.kitab
Overview
🕮 KITAB is a challenging dataset and a dynamic data collection approach for testing abilities of Large Language Models (LLMs) in answering information retrieval queries with constraint filters. A filtering query with constraints can be of the form "List all books written by Toni Morrison that were published between 1970-1980". The dataset was originally contributed by the paper "KITAB: Evaluating LLMs on Constraint Satisfaction for Information Retrieval" Marah I Abdin… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/kitab.bing_coronavirus_query_set
Dataset Card for BingCoronavirusQuerySet
Dataset Summary
Please note that you can specify the start and end date of the data. You can get start and end dates from here: https://github.com/microsoft/BingCoronavirusQuerySet/tree/master/data/2020
example:
load_dataset("bing_coronavirus_query_set", queries_by="state", start_date="2020-09-01", end_date="2020-09-30")
You can also load the data by country by using queries_by="country".
Supported Tasks and… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/bing_coronavirus_query_set.MuseVLA-dataset
MuseVLA Dataset
Multi-modal robot manipulation dataset with synchronized RGB, depth, acoustic,
thermal, and radar streams. Released as two parts (dataset_01/,
dataset_02/) sharing the same per-episode layout. Together they cover
~1400 episodes across 11 instructions (towel / clothes / box / item / drink
manipulation).
Per-episode contents
{episode_name}/
├── video.mp4 # RGB, 1280×720, 30 fps
├── mask/video.mp4 #… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/MuseVLA-dataset.msr-acc-tae25
Microsoft Research - Accurate Chemistry Collection: Total Atomization Energies
Description
The Microsoft Research Accurate Chemistry Collection (MSR-ACC) provides a collection of accurate coupled cluster labels for training machine learning functionals.
MSR-ACC/TAE25 comprising 73,040 total atomization energies at the CCSD(T)/CBS level obtained with the W1-F12 thermochemical protocol.
The dataset is constructed to exhaustively cover the chemical space of closed-shell… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/msr-acc-tae25.mediflow
MediFlow
A large-scale synthetic instruction dataset of 2.5M rows (~700k unique instructions) for clinical natural language processing covering 14 task types and 98 fine-grained input clinical documents.
t-SNE 2D Plot of MediFlow Embeddings by Task Types
Dataset Splits
mediflow: 2.5M instruction data for SFT alignment.
mediflow_dpo: ~135k top-quality instructions with GPT-4o generated rejected_output for DPO alignment.
Main Columns
instruction:… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/mediflow.WildFeedback
Dataset Card for WildFeedback
WildFeedback is a preference dataset constructed from real-world user interactions with ChatGPT. Unlike synthetic datasets that rely solely on AI-generated rankings, WildFeedback captures authentic human preferences through naturally occurring user feedback signals in conversation. The dataset is designed to improve the alignment of large language models (LLMs) with actual human values by leveraging direct user input.
Dataset Details… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/WildFeedback.WorkflowPerturb
WorkflowPerturb — Dataset Artifact
Companion data for the EMNLP 2026 Industry Track paper
“WorkflowPerturb: Calibrated Stress Tests for Evaluating Multi-Agent Workflow Metrics.”
Canonical location: https://huggingface.co/datasets/microsoft/WorkflowPerturbPaper: https://arxiv.org/abs/2602.17990
This release is the complete WorkflowPerturb benchmark plus documentation. It is
self-contained: the CSVs carry every golden workflow, every perturbed variant, and all
shipped pre-computed… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/WorkflowPerturb.benchpress-score-matrix
BenchPress Score Matrix
This dataset contains the public model-by-benchmark score matrix used by
BenchPress. The release includes the lossless audited JSON, benchmark cost
evidence, flat model and benchmark metadata, one row per observed score, and
the paper-canonical dense subset used in the BenchPress experiments.
The source repository is
microsoft/benchpress.
Canonical artifacts
data/llm_benchmark_data.json is the authoritative rich score-matrix artifact.
It… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/benchpress-score-matrix.PatientSafetyBench
Disclaimer
The synthetic prompts may contain offensive, discriminatory, or harmful language. These fake prompts also mention topics that are not based on the scientific consensus at all.These prompts are included solely for the purpose of evaluating safety behavior of language models.
⚠️ Disclaimer: The presence of such prompts does not reflect the views, values, or positions of the authors, their institutions, or any affiliated organizations. They are provided exclusively for… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/PatientSafetyBench.hnm-search-data
HnM Search Dataset Created from Recommendations Dataset
This synthetic data-set is created using the recommendations dataset:
https://huggingface.co/datasets/einrafh/hnm-fashion-recommendations-data (Use of this dataset is subject to the terms and conditions set forth on the original distribution page. This dataset is intended for non-commercial and research use.)
https://www.kaggle.com/competitions/h-and-m-personalized-fashion-recommendations/data (DATA ACCESS AND USE:… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/hnm-search-data.CoSAlign-Train
CoSAlign-Train: A Large-Scale Synthetic Training Dataset for Controllable Safety Alignment
Paper: Controllable Safety Alignment: Inference-Time Adaptation to Diverse Safety Requirements, published at ICLR 2025.
Purpose: Training dataset for controllable safety alignment (CoSA) of large language models (LLMs), facilitating fine-grained inference-time adaptation to diverse safety requirements.
Description: CoSAlign-Train is a large-scale, synthetic preference dataset designed for… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/CoSAlign-Train.tspcleand_microsoft_rStar-Coder元データ: https://huggingface.co/datasets/microsoft/rStar-Coder
データ件数: 269,863
平均トークン数: 11674
最大トークン数: 31,184
合計トークン数: 3,150,447,484
ファイル形式: JSONL
ファイルサイズ: 不明
加工内容
synthetic_sftを使用
トークン処理が重たいので、文字数でフィルター
seed_question < 6000
generation < 80000
thinkタグ除去 が中途半端なものを除外
トークナイズ処理（速度向上アップデート
繰り返し除去
microsoft-Phi-3-mini-4k-instruct-toksuite-detokenizedTraining data of the model detokenized in the exact order seen by the model.
The training data is partitioned into 8 chunks (chunk-0 through chunk-7), based on the GPU rank that generated the data. Each chunk contains detokenized text files in JSON Lines format (.jsonl).
microsoft__phi-4-details
Dataset Card for Evaluation run of microsoft/phi-4
Dataset automatically created during the evaluation run of model microsoft/phi-4
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 3 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional configuration… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__phi-4-details.satmicrosoft__phi-2-details
Dataset Card for Evaluation run of microsoft/phi-2
Dataset automatically created during the evaluation run of model microsoft/phi-2
The dataset is composed of 44 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional configuration… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__phi-2-details.microsoft__Phi-3-mini-4k-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3-mini-4k-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3-mini-4k-instruct
The dataset is composed of 73 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 6 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3-mini-4k-instruct-details.microsoft__phi-1_5-details
Dataset Card for Evaluation run of microsoft/phi-1_5
Dataset automatically created during the evaluation run of model microsoft/phi-1_5
The dataset is composed of 44 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__phi-1_5-details.microsoft__Phi-3-medium-4k-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3-medium-4k-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3-medium-4k-instruct
The dataset is composed of 44 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3-medium-4k-instruct-details.microsoft__Phi-3-mini-128k-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3-mini-128k-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3-mini-128k-instruct
The dataset is composed of 40 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3-mini-128k-instruct-details.reddit_finance_posts_apple-tesla-microsoft
Reddit Finance Posts Dataset (Apple, Tesla, Microsoft)
This dataset contains 12046 Reddit Posts collected from 20 finance-related subreddits via the Reddit API using the keywords Apple, Tesla, and Microsoft.
The included subreddits are:
stocks, wallstreetbets, investing, StockMarket, options, RobinHood,
pennystocks, SecurityAnalysis, personalfinance, Dividends, CryptoCurrency,
CryptoMarkets, ETFs, FinancialIndependence, ValueInvesting, quant,
algotrading, forex, economy, Superstonk… See the full description on the dataset page: https://huggingface.co/datasets/emilpartow/reddit_finance_posts_apple-tesla-microsoft.microsoft__Phi-4-mini-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-4-mini-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-4-mini-instruct
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-4-mini-instruct-details.microsoft__Phi-3.5-mini-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3.5-mini-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3.5-mini-instruct
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3.5-mini-instruct-details.microsoft__Phi-3.5-MoE-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3.5-MoE-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3.5-MoE-instruct
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3.5-MoE-instruct-details.microsoft__Phi-3-medium-128k-instruct-details
Dataset Card for Evaluation run of microsoft/Phi-3-medium-128k-instruct
Dataset automatically created during the evaluation run of model microsoft/Phi-3-medium-128k-instruct
The dataset is composed of 38 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 1 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__Phi-3-medium-128k-instruct-details.microsoft__phi-1-details
Dataset Card for Evaluation run of microsoft/phi-1
Dataset automatically created during the evaluation run of model microsoft/phi-1
The dataset is composed of 44 configuration(s), each one corresponding to one of the evaluated task.
The dataset has been created from 2 run(s). Each run can be found as a specific split in each configuration, the split being named using the timestamp of the run.The "train" split is always pointing to the latest results.
An additional configuration… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard/microsoft__phi-1-details.