datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
answercarefully-dpo-ja-2026
AnswerCarefully-derived Japanese DPO data for LLM safety
本データセットは、llm-jp/AnswerCarefullyを参照して作成した日本語LLMの安全応答をDPOで学習するためのpreference datasetです。
利用条件
本データセットには、llm-jp/AnswerCarefullyと同じ利用規約を適用します。
利用者は、llm-jp/AnswerCarefullyと本データセットの両方で利用規約に同意する必要があります。
データ
train: 417件
validation: 44件
各行には次のフィールドが含まれます。
id: 本リリース内だけで使用するID
prompt: 元質問の意味と危険性を変えずに言い換えた質問
chosen: DPOで望ましい応答として扱う回答
rejected: DPOで望ましくない応答として扱う回答
category, harm_type, risk_area… See the full description on the dataset page: https://huggingface.co/datasets/ekunish/answercarefully-dpo-ja-2026.rlvr-reward-hacking-scale-no-conftest-20260909-completion
Matched no-conftest RLVR study 20260909-completion
Lossless research records, grouped by model and trajectory type. Only the listed
configurations have published records. Canary diagnostics are excluded from study
estimates; run status in provenance distinguishes retired diagnostics from active
or completed training. Valid failures, refusals and truncations are retained.
The train split name is a dataset-loader convention; record_type identifies
whether a record is training… See the full description on the dataset page: https://huggingface.co/datasets/lucabaroni/rlvr-reward-hacking-scale-no-conftest-20260909-completion.RoadmapBench
RoadmapBench
A benchmark for evaluating AI coding agents on multi-target, long-horizon software development tasks derived from open-source project version upgrades.
Overview
RoadmapBench contains 115 tasks spanning 17 open-source repositories across 5 programming languages (Python, TypeScript, Go, Rust, C++). Each task requires an agent to implement multiple interdependent features that correspond to a real version upgrade of the target project.
Task Structure… See the full description on the dataset page: https://huggingface.co/datasets/benchmark-anon-2026/RoadmapBench.MatrAIx_Persona_1M
MatrAIx Persona 1M
999,847 personas, each described by 1,290 categorical attributes.
599,847 are derived from real records, 400,000 are synthetic.
10 Zstandard Parquet shards, 4.17 GB.
Read it with pyarrow, not datasets
Attributes are packed: one persona's 1,290 attributes are 645 bytes of 4-bit
codes, low nibble first. datasets cannot open these files at all. Use pyarrow
and decode against persona_codes.schema.json.
import json, pyarrow.parquet as pq
schema =… See the full description on the dataset page: https://huggingface.co/datasets/MatrAIx2026/MatrAIx_Persona_1M.HRM-Text-data-io-cleaned-20260515Pre-built HRM-Text pretraining dataset from raw data using the data_io cleaning scripts.
Citation
If you find this project or our paper useful, please consider citing our paper:
@misc{wang2026hrmtextefficientpretrainingscaling,
title={HRM-Text: Efficient Pretraining Beyond Scaling},
author={Guan Wang and Changling Liu and Chenyu Wang and Cai Zhou and Yuhao Sun and Yifei Wu and Shuai Zhen and Luca Scimeca and Yasin Abbasi Yadkori},
year={2026}… See the full description on the dataset page: https://huggingface.co/datasets/sapientinc/HRM-Text-data-io-cleaned-20260515.US_Regulation_ECFR_20260101
US Regulation eCFR 2026-01-01 Dataset
This repository contains a structured, machine-readable version of the Electronic Code of Federal Regulations (eCFR), captured as of January 1, 2026. Unlike the annual CFR snapshots, this dataset reflects the editorialized, near real-time version of federal regulations.
Dataset Description
The eCFR is a daily updated editorial compilation of CFR material and Federal Register amendments. This dataset captures a specific point-in-time… See the full description on the dataset page: https://huggingface.co/datasets/Azzindani/US_Regulation_ECFR_20260101.US_Regulation_CFR_20260101stack-v3-train
🥞 The Stack v3
What is it?
What is being released
How to download and use it
Dataset statistics
Dataset structure
Dataset creation
Considerations for using the data
Additional information
What is it?
The Stack v3 is the largest, most up-to-date open dataset of source code, crawled directly from GitHub and built to pre-train code LLMs with full-repository context. It is the successor to The Stack v2 and, like its predecessor, is released to make the training… See the full description on the dataset page: https://huggingface.co/datasets/happahhap2026/stack-v3-train.2026.RA.Negotiation-Campaigns
Rational-Agent Negotiation Campaigns
This public dataset contains the complete selected evidence for the
ii_mats/experiments/rational_agents negotiation experiments. It includes raw
episode JSON, post-hoc annotations, Markdown and HTML transcripts, committed
instances, run manifests, campaign selection and exclusion ledgers,
machine-readable analysis tables, figures, and integrity manifests.
No contaminated, duplicated, stale, failed, or superseded run is included as
selected… See the full description on the dataset page: https://huggingface.co/datasets/siddharthmb/2026.RA.Negotiation-Campaigns.AIME_2000_2026_Kimi_K3
AIME 2000–2026 — Kimi K3 reasoning traces
🔄 Changelog
2026-08-08 — full re-generation. All reasoning traces were regenerated from scratch and re-verified against the official answer key.
New schema — added gen_attempts_low, gen_attempts_high; renamed gen_parsed_answer → gen_answer_int and answer_note → problem_note; removed gen_effort, gen_pass1.
New generation — only use the bare problem (v1 appended an "ANSWER:" format instruction), so traces are cleaner.… See the full description on the dataset page: https://huggingface.co/datasets/bevangelista/AIME_2000_2026_Kimi_K3.general-master-en-202608
General · Master · English · 2026-08
English pretraining text, assembled from three public sources, cleaned with one
character-level cleaner, and filtered for repetition.
109,337,531 documents and 468,064,046,462 characters.
Composition
Config
Documents
Characters
What it is
fineweb-edu-dedup
65,010,430
297,544,916,118
Web text an educational classifier kept
cosmopedia-v2
38,591,146
144,011,993,012
Synthetic prose from a seeded generator… See the full description on the dataset page: https://huggingface.co/datasets/Brainquiver/general-master-en-202608.Wiki_FR_2026.07_TexteIntroductif
Version Complète : MisterAI/WM-ENT-API-DUMP_FR_2026.07
https://huggingface.co/datasets/MisterAI/WM-ENT-API-DUMP_FR_2026.07
ESSAI I : Section Introductive Uniquement :: Jeux De Données : Dump WikiMedia Français Juillet 2026 : Extraction et Nettoyage
Description
Ce JDD contient des articles extraits du dump complet de Wikimedia Enterprise de juillet 2026, nettoyés et structurés pour l'entraînement de modèles d'apprentissage automatique.
Source… See the full description on the dataset page: https://huggingface.co/datasets/MisterAI/Wiki_FR_2026.07_TexteIntroductif.rlvr-reward-hacking-scale-no-conftest-20260909
Matched no-conftest RLVR study 20260909
Complete immutable training, monitoring and comparison trajectories for six models.
All valid outcomes are retained, including refusals, failures and truncations.
The train split name is a dataset-loader convention; record_type identifies
whether a record is training, monitoring, comparison, or a derived judgment.
import json
from datasets import load_dataset
rows = load_dataset("lucabaroni/rlvr-reward-hacking-scale-no-conftest-20260909"… See the full description on the dataset page: https://huggingface.co/datasets/lucabaroni/rlvr-reward-hacking-scale-no-conftest-20260909.general-web-it-202608
General · Web · Italian · 2026-08
Italian pretraining text, built from the Italian portion of EPFL's FineWeb2-HQ, which is the
high quality slice of FineWeb-2. Every document passes one character-level cleaner and a
repetition filter.
21,065,052 documents and 66,158,573,443 characters of Italian prose.
Contents
Config
Documents
Characters
Upstream
fineweb2-hq-ita_Latn
21,065,052
66,158,573,443
epfml/FineWeb2-HQ, ita_Latn
The character count is… See the full description on the dataset page: https://huggingface.co/datasets/Brainquiver/general-web-it-202608.general-web-fr-202608
General · Web · French · 2026-08
French pretraining text, built from the French portion of EPFL's FineWeb2-HQ, which is the
high quality slice of FineWeb-2. Every document passes one character-level cleaner and a
repetition filter.
31,999,309 documents and 118,346,333,763 characters of French prose.
Contents
Config
Documents
Characters
Upstream
fineweb2-hq-fra_Latn
31,999,309
118,346,333,763
epfml/FineWeb2-HQ, fra_Latn
The character count is exact.… See the full description on the dataset page: https://huggingface.co/datasets/Brainquiver/general-web-fr-202608.simverse2026
SimVerse
⚠️ Anonymized for double-blind review. This dataset is currently undergoing peer review. It is hosted under an anonymous account dedicated to the review process; the author and citation fields are deliberately unfilled. Permanent ownership and citation information will be added after the review concludes. Please do not attempt to deanonymize the maintainers of this dataset during review.
A multi-task benchmark for evaluating multimodal LLMs on interactive simulation… See the full description on the dataset page: https://huggingface.co/datasets/SimVer-ano/simverse2026.meddialbench
MedDialBench
A controlled factorial benchmark for evaluating LLM diagnostic robustness under parametric adversarial patient behaviors.
Anonymous submission to NeurIPS 2026 Datasets and Benchmarks Track.
The companion paper is currently under double-blind review. Author identities and institutional affiliations are intentionally omitted. After publication, this repository will be transferred to a permanent (de-anonymized) location and this README updated accordingly.… See the full description on the dataset page: https://huggingface.co/datasets/anon-meddial-2026/meddialbench.zebra-cot-mistral-small-3.2-24b-preprocessed
Zebra-CoT Preprocessed — Mistral Hackathon 2026
Preprocessed version of the Zebra-CoT dataset for fine-tuning Mistral-Small-3.2-24B-Instruct.
Format
text: formatted as [INST] question [/INST] <think> reasoning </think> answer
image: PIL JPEG image for the corresponding visual task
Usage
Fine-tuning Mistral-Small-3.2-24B on chain-of-thought visual reasoning.
Hackathon
Created for Mistral Hackaton 2026 — Fine-tuning track with W&B.
HRM-Text-data-io-cleaned-20260515-copyPre-built HRM-Text pretraining dataset from raw data using the data_io cleaning scripts.
Citation
If you find this project or our paper useful, please consider citing our paper:
@misc{wang2026hrmtextefficientpretrainingscaling,
title={HRM-Text: Efficient Pretraining Beyond Scaling},
author={Guan Wang and Changling Liu and Chenyu Wang and Cai Zhou and Yuhao Sun and Yifei Wu and Shuai Zhen and Luca Scimeca and Yasin Abbasi Yadkori},
year={2026}… See the full description on the dataset page: https://huggingface.co/datasets/huankguan2/HRM-Text-data-io-cleaned-20260515-copy.terminal_bench_2_tasktrove_dq_unitsyn_python_step20_30b_a3b_20260730_014827
TaskTrove DQ unitsyn-python training traces (step 20, 30B-A3B)
Terminus-2 agent rollouts recorded while training
laion/tasktrove-dq-unitsyn-python-step20-30b-a3b
with SkyRL from Qwen/Qwen3-Coder-30B-A3B-Instruct.
Each row is the last episode of one trial: the full agent transcript, the task instruction, the
scalar reward, and the verifier's output.
Source run: rl-tasktrove-dq-sweep-30b-terminus2-qwen-20260725-163115-1ae770.
Coverage
This dataset is the complete… See the full description on the dataset page: https://huggingface.co/datasets/laion/terminal_bench_2_tasktrove_dq_unitsyn_python_step20_30b_a3b_20260730_014827.groundwork-tech-2026
Groundwork Tech 2026
Open dataset for Groundwork tech pillar — 25 articles.
Source: https://gworky.com/tech
See data.json for records.
groundwork-money-2026
Groundwork Money 2026
Open dataset for Groundwork money pillar — 25 articles.
Source: https://gworky.com/money
See data.json for records.
cybersecurity-soc-threat-hunting-sft-dpo-2026
🛡️ Enterprise Cybersecurity AI, SOC Tier-3 & Threat Hunting SFT/DPO Dataset (2026)
High-precision multi-turn instruction tuning and preference optimization dataset with step-by-step SOC Tier-3 Chain-of-Thought (<thought>) kill-chain diagnostic trees for fine-tuning LLMs (Llama-3.3, Qwen-2.5-Coder, DeepSeek-R1-Distill, Mistral) into Senior SOC Threat Hunters, Incident Responders, and Red-Team Defense Architects.
📊 Dataset Architecture & Highlights… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/cybersecurity-soc-threat-hunting-sft-dpo-2026.recube-dataThis dataset serves as the official data repository for the ReCUBE benchmark; the corresponding evaluation codebase and execution instructions can be found at https://anonymous.4open.science/r/ReCUBE-E0FB/README.md.
Data
This directory contains all benchmark data for ReCUBE.
Download
All data files are hosted on Hugging Face and can be downloaded using:
# Install huggingface_hub if not already installed
pip install huggingface_hub
# Download the entire dataset… See the full description on the dataset page: https://huggingface.co/datasets/recube-anon-2026/recube-data.danbooru-wiki-2026
danbooru-wiki-2026-04-28
About
Wiki pages about the danbooru tags on danbooru.donmai.us. The wiki contains the description of each tag.
This dataset was collected from the wiki data using danbooru API and then merged with tags data (category_name & post_count). No entries were removed, no matter how many posts they have. The resulting set was manually filtered. Originally some of the lists, tag_group:, about:, help:, howto:, etc, pages were labled as "general"… See the full description on the dataset page: https://huggingface.co/datasets/kierarkia/danbooru-wiki-2026.financial-statement-modeling-sft-dpo-2026
📈 Enterprise Financial AI, SEC 10-K & Valuation Modeling SFT/DPO Dataset (2026)
High-precision multi-turn instruction tuning and preference optimization dataset with step-by-step arithmetic Chain-of-Thought (<thought>) reasoning chains for fine-tuning LLMs (Llama-3.3, Qwen-2.5-Coder, DeepSeek-R1-Distill, Mistral) into Wall Street Equity Research Associates, M&A Valuation Modelers, and Senior Forensic Auditors.
📊 Dataset Architecture & Highlights
Multi-Turn… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/financial-statement-modeling-sft-dpo-2026.groundwork-life-2026
Groundwork Life 2026
Open dataset for Groundwork life pillar — 25 articles.
Source: https://gworky.com/life
See data.json for records.
groundwork-home-2026
Groundwork Home 2026
Open dataset for Groundwork home pillar — 25 articles.
Source: https://gworky.com/home
See data.json for records.
groundwork-body-2026
Groundwork Body 2026
Open dataset for Groundwork body pillar — 25 articles.
Source: https://gworky.com/body
See data.json for records.
swe-bench-multi-file-refactoring-sft-dpo-2026
💻 Enterprise Autonomous SWE-bench AI & Multi-File Code Refactoring SFT/DPO Dataset (2026)
High-precision multi-turn instruction tuning and preference optimization dataset with step-by-step call-stack Chain-of-Thought (<thought>) reasoning trees for fine-tuning LLMs (Qwen-2.5-Coder, Llama-3.3, DeepSeek-R1-Distill, Mistral) into Autonomous Software Engineers and SWE-bench Benchmark Agents.
📊 Dataset Architecture & Highlights
Multi-Turn Code Reviews:… See the full description on the dataset page: https://huggingface.co/datasets/beatsprom/swe-bench-multi-file-refactoring-sft-dpo-2026.