datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Orchard
Orchard Dataset
Overview
Orchard is the trajectory release accompanying the paper "Orchard: An Open-Source Agentic Modeling Framework" (Peng et al., 2026). It bundles two parallel agentic-modeling datasets distilled from strong teacher models, both produced inside the same Orchard Env sandbox infrastructure:
swe — 107,185 multi-turn software-engineering trajectories across 2,788 GitHub repositories, each labeled with whether the agent's final patch passed the… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/Orchard.OpenMementos
OpenMementos-228K
A dataset of 228,557 reasoning traces annotated with block segmentation and compressed summaries (mementos), derived from OpenThoughts-v3.
Memento is a framework for teaching language models to manage their own context during long-form reasoning. Instead of generating one long, unstructured chain-of-thought, memento-trained models segment their reasoning into blocks, compress each block into a dense summary (a memento), and continue reasoning from mementos alone.… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/OpenMementos.delegate52
DELEGATE52
Overview
DELEGATE52 is a benchmark dataset for evaluating LLMs on long-horizon delegated document editing across 52 professional document domains (crystallography files, music notation, accounting ledgers, Python source code, etc.). The dataset was developed to study the readiness of AI systems for delegated workflows, a new interaction paradigm where knowledge workers instruct LLMs to edit documents on their behalf over long sessions.
A detailed… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/delegate52.delulu-fim-benchmarkDelulu — Fill-in-the-Middle Code Hallucination Benchmark
A verified multilingual benchmark for code-completion hallucinations.
Every golden completion compiles. Every hallucination provably doesn't.
📄 Read the preprint on arXiv →
Every Delulu sample ships as a self-contained Docker image. The viewer above lets you browse the dataset, pull a sample's verifier, and re-run verify golden / verify hallucinated / verify patch <your-completion> with one… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/delulu-fim-benchmark.NextCoderDataset
NextCoderDataset
GitHub | Paper
NextCoder: Robust Adaptation of Code LMs to Diverse Code Edits (ICML'2025)
Data Overview
NextCoderdataset is the instruction-variant of synthetic dataset, used for training models on code-editing scenarios and compromised of around 381k (127k*3) samples across 8 different programming languages: Python,
Java, C++, C, Rust, Javascript, Go and Kotlin.
This is used to finetune the NextCoder family models using the novel… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/NextCoderDataset.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.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.FStarDataSet
Proof Oriented Programming with AI (PoPAI) - FStarDataSet
This dataset contains programs and proofs in F* proof-oriented programming language.
The data, proposed in Towards Neural Synthesis for SMT-Assisted Proof-Oriented Programming,
is an archive of source code, build artifacts, and metadata assembled from eight different F⋆-based open source projects on GitHub.
Primary-Objective
This dataset's primary objective is to train and evaluate Proof-oriented Programming… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/FStarDataSet.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.FStarDataSet-V2This dataset is the Version 2.0 of microsoft/FStarDataSet.
Primary-Objective
This dataset's primary objective is to train and evaluate Proof-oriented Programming with AI (PoPAI, in short). Given a specification of a program and proof in F*,
the objective of a AI model is to synthesize the implemantation (see below for details about the usage of this dataset, including the input and output).
Data Format
Each of the examples in this dataset are organized as dictionaries… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/FStarDataSet-V2.openpii-masking-micro-100k
OpenPII Micro: Multilingual PII Masking Sample
A micro-sized stratified sample of OpenPII 1.5M,
perfect for quick prototyping, smoke tests, and CI fixtures. Every locale and every
label that exists in the parent dataset is represented in proportion.
📖 More information: www.ai4privacy.com/datasets/pii-masking-3m-asia-pacific
Dataset Details
Total Examples
Train
Validation
Labels
Languages
Regions
Annotations
Format
License
100,000
90,000
10,000
19… See the full description on the dataset page: https://huggingface.co/datasets/ai4privacy/openpii-masking-micro-100k.OB-Inference-Microtasks
Inference Microtasks
29 synthetic microtasks with reference answers across meeting-notes lookup,
support-ticket triage, and contract-terms extraction. The public set accompanies
the OpenBenchmarks Inference Benchmark,
which measures single-user delay on short, deliberately easy structured tasks.
Dataset contents
Configuration
Rows
Task
contract-terms-extraction
10
Extract commercial terms from a technology contract excerpt.
meeting-notes-lookup
13… See the full description on the dataset page: https://huggingface.co/datasets/openbenchmarks/OB-Inference-Microtasks.pii-masking-micro-100k
PII Masking Micro: Multilingual Sample
A micro-sized stratified sample of pii-masking-openpii-1.5m,
the flagship release of the PII-Masking-3M family. Sampled proportionally by
(source_dataset, language) so every locale and label gets representation.
Asia Pacific rows appear first.
📖 More information: www.ai4privacy.com/datasets/pii-masking-3m-asia-pacific
Dataset Details
Total Examples
Train
Validation
Labels
Languages
Regions
Annotations
Format… See the full description on the dataset page: https://huggingface.co/datasets/ai4privacy/pii-masking-micro-100k.RealStories-Micro-MRL
Dataset Card for ReactiveAI/RealStories-Micro-MRL
First synthetic Memory Reinforcement Learning dataset for Proof-of-Concept Reactive Transformer models.
Dataset is divided into subsets, used in different Curriculum Stage of MRL training - each subset have
different number of follow-up interactions, could use different strategy, and have train and validation
splits.
Subsets
steps-1: ~2300 train (~4600 interactions) / ~340 validation (~680 interactions) - Single-Step… See the full description on the dataset page: https://huggingface.co/datasets/ReactiveAI/RealStories-Micro-MRL.SYNUR
Dataset Card: SYNUR (Synthetic Nursing Observation Dataset)
1. Dataset Summary
Name: SYNUR
Full name / acronym: SYnthetic NURsing Observation Extraction
Purpose / use case:SYNUR is intended to support research in structuring nurse dictation transcripts by extracting clinical observations that can feed into flowsheet-style EHR entries.
It is designed to reduce documentation burden by enabling automated conversion from spoken nurse assessments to structured… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/SYNUR.unpredictable_msdn-microsoft-comThe UnpredicTable dataset consists of web tables formatted as few-shot tasks for fine-tuning language models to improve their few-shot performance. For more details please see the accompanying dataset card.prototypical-hai-collaborationsPaper: Prototypical Human-AI Collaboration Behaviors from LLM-Assisted Writing in the Wild
LICENSE: ODC-BY
Contact: Sheshera Mysore, Bahar Sarrafzadeh
Introduction
The repository releases code and data for the paper: Prototypical Human-AI Collaboration Behaviors from LLM-Assisted Writing in the Wild. The dataset release only contains the public WildChat-1M dataset annotated with labels used for the analysis in the paper.
Dataset contents
wildchat1m_en3u-task_anns.jsonl:… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/prototypical-hai-collaborations.reasoning-sft-minimax-microsoft-orca-agentinstruct-1M-v1
MiniMax-M2.5 Reasoning SFT (Orca AgentInstruct 1M v1)
Reasoning SFT dataset generated by MiniMaxAI/MiniMax-M2.5 on prompts from the Stratified K-Means Diverse Instruction-Following 100K-1M dataset (Orca AgentInstruct subset).
Format
Each row has three columns:
input — list of dicts [{"role": "...", "content": "..."}, ...] (conversation turns)
response — model-generated response with <think> reasoning block
source — task category (creative_content, text_modification, rc… See the full description on the dataset page: https://huggingface.co/datasets/AmanPriyanshu/reasoning-sft-minimax-microsoft-orca-agentinstruct-1M-v1.task707_mmmlu_answer_generation_high_school_microeconomics
Dataset Card for Natural Instructions (https://github.com/allenai/natural-instructions) Task: task707_mmmlu_answer_generation_high_school_microeconomics
Additional Information
Citation Information
The following paper introduces the corpus in detail. If you use the corpus in published work, please cite it:
@misc{wang2022supernaturalinstructionsgeneralizationdeclarativeinstructions,
title={Super-NaturalInstructions: Generalization via Declarative… See the full description on the dataset page: https://huggingface.co/datasets/Lots-of-LoRAs/task707_mmmlu_answer_generation_high_school_microeconomics.MM-WebGen-Bench
MM-WebGen-Bench: A Benchmark for Multimodal Webpage Generation
MM-WebGen-Bench is a multi-level evaluation benchmark for multimodal webpage generation, proposed in MM-WebAgent. It contains 120 curated webpage design prompts covering 11 scene categories, 11 visual styles, and diverse multimodal compositions (4 video types, 8 image types, and 17 chart types).
Links
Project Page: aka.ms/mm-webagent
GitHub: microsoft/MM-webagent
Paper: MM-WebAgent: A Hierarchical… See the full description on the dataset page: https://huggingface.co/datasets/microsoft/MM-WebGen-Bench.offline-micro-saas-catalog
📦 SaveDollars.store — Offline Micro SaaS & Autonomous AI Software Catalog
This dataset contains structured product metadata, architecture specifications, pricing, and documentation for 96 standalone offline Micro SaaS applications, autonomous AI agent command centers, and business operating systems published by SaveDollars.store.
📊 Dataset Structure (catalog.json)
Each record represents a production-ready, subscription-free software package:
{
"id": 75809… See the full description on the dataset page: https://huggingface.co/datasets/SaveDollars/offline-micro-saas-catalog.BCE-Prettybird-Micro-Standard-v0.0.1
🚀 The Future Standard / Geleceğin Standartı
[English]
Beyond Raw Data: The Behavioral Revolution
The AI industry has been obsessed with the volume of data. At Prometech A.Ş., we are shifting the focus to the process of thought. BCE-Prettybird-Micro-Standart is not just a collection of Q&As; it is a blueprint for behavioral reasoning. By integrating Path Mapping and Behavioral DNA into the training loop, we are setting the new industry standard: Small models with elite… See the full description on the dataset page: https://huggingface.co/datasets/pthinc/BCE-Prettybird-Micro-Standard-v0.0.1.microagent-train-v3
microagent-train-v3
SFT corpus for training a 4B-class terminal-agent model (Qwen3-4B-Thinking) in the microagent XML protocol. v3 = the 26,627-trajectory v2 corpus plus 3,951 synthetic failure-recovery trajectories that target the specific execution weaknesses found in v1 evaluation.
Why v3 exists
The v1 model (prometheus04/qwen3-4b-thinking-microagent-v1-merged) scored 1/89 (1.12%) on Terminal-Bench 2.0. Trajectory analysis showed the model reasoned correctly but failed… See the full description on the dataset page: https://huggingface.co/datasets/prometheus04/microagent-train-v3.obekt-question-answer-reasoning-micro-v0.1
Obekt Micro Reasoning Dataset (v0.1)
Dataset Description
This is a "micro" dataset containing questions, answers, and reasoning traces. It is generated using the Xiaomi MiMo V2 Flash LLM and is intended for experimental purposes, quick prototyping, and fine-tuning trials where reasoning capability is a focus.
Source Model: xiaomi/mimo-v2-flash
Contains
obekt-question-answer-reasoning-micro-v0.1.csv: The main data file.
Columns:
question: The input query.… See the full description on the dataset page: https://huggingface.co/datasets/obekt/obekt-question-answer-reasoning-micro-v0.1.BCE-Prettybird-Micro-Standard-v0.0.2
🚀 The Future Standard / Geleceğin Standartı
[English]
Beyond Raw Data: The Behavioral Revolution
The AI industry has been obsessed with the volume of data. At Prometech A.Ş., we are shifting the focus to the process of thought. BCE-Prettybird-Micro-Standart is not just a collection of Q&As; it is a blueprint for behavioral reasoning. By integrating Path Mapping and Behavioral DNA into the training loop, we are setting the new industry standard: Small models with elite… See the full description on the dataset page: https://huggingface.co/datasets/pthinc/BCE-Prettybird-Micro-Standard-v0.0.2.Vietnamese-microsoft-orca-math-word-problems-200k-gg-translatedBurmese-Microbiology-1K
Burmese-Microbiology-1K
Min Si Thu, min@globalmagicko.com
Microbiology 1K QA pairs in Burmese Language
Purpose
Before this Burmese Clinical Microbiology 1K dataset, the open-source resources to train the Burmese Large Language Model in Medical fields were rare.
Thus, the high-quality dataset needs to be curated to cover medical knowledge for the development of LLM in the Burmese language
Motivation
I found an old notebook in my box. The book was… See the full description on the dataset page: https://huggingface.co/datasets/jojo-ai-mst/Burmese-Microbiology-1K.BCE-Prettybird-Micro-Standard-v0.0.4
🚀 The Future Standard / Geleceğin Standartı
[English]
Beyond Raw Data: The Behavioral Revolution
The AI industry has been obsessed with the volume of data. At Prometech A.Ş., we are shifting the focus to the process of thought. BCE-Prettybird-Micro-Standart is not just a collection of Q&As; it is a blueprint for behavioral reasoning. By integrating Path Mapping and Behavioral DNA into the training loop, we are setting the new industry standard: Small models with elite… See the full description on the dataset page: https://huggingface.co/datasets/pthinc/BCE-Prettybird-Micro-Standard-v0.0.4.BCE-Prettybird-Micro-Standard-v0.0.3
🚀 The Future Standard / Geleceğin Standartı
[English]
Beyond Raw Data: The Behavioral Revolution
The AI industry has been obsessed with the volume of data. At Prometech A.Ş., we are shifting the focus to the process of thought. BCE-Prettybird-Micro-Standart is not just a collection of Q&As; it is a blueprint for behavioral reasoning. By integrating Path Mapping and Behavioral DNA into the training loop, we are setting the new industry standard: Small models with elite… See the full description on the dataset page: https://huggingface.co/datasets/pthinc/BCE-Prettybird-Micro-Standard-v0.0.3.