Independent project. Not a U.S. government website.Independent project. Not a United States government website.
Support our projectSupport usSupport our project · Buy me a coffee (opens Buy Me a Coffee)Interactive Search Tool: (paper) License: ODC-BY Language(s) (NLP): multi-lingual Point of Contact: Yuntian Deng Dataset Summary WildChat is a collection of 1 million co…
474 likes36,853 downloads in the last 30 daysText generationODC-By
We propose the Chinese Open Instruction Generalist (COIG) project to maintain a harmless, helpful, and diverse set of Chinese instruction corpora. We welcome all researc…
470 likes1,713 downloads in the last 30 daysApache 2.0
This dataset comprises high-quality audio recordings of a standardized 50-word paragraph, designed for training and evaluating automatic speech recognition (ASR) models.…
467 likes234 downloads in the last 30 daysNo license given
This dataset contains a collection of street videos designed to support tasks such as video classification, object detection, and urban scene analysis. Each video captur…
466 likes350 downloads in the last 30 daysVideo classificationCC BY 4.0
*This dataset contains intentionally low-quality (“B-grade”) data. It has been curated to include noisy, imperfect, or otherwise suboptimal samples for the purpose of te…
465 likes50 downloads in the last 30 daysAudio classificationCC BY 4.0
HelpSteer2: Open-source dataset for training top-performing reward models HelpSteer2 is an open-source Helpfulness Dataset (CC-BY-4.0) that supports aligning models to b…
465 likes32,805 downloads in the last 30 daysCC BY 4.0
Ultra-FineWeb is a large-scale, high-quality, and efficiently-filtered dataset. We use the proposed efficient verification-based high-quality filtering pipeline to the F…
457 likes65,634 downloads in the last 30 daysText generationApache 2.0
Ended up with some tokens to burn on a Claude Max plan. Assembly began during 4.6 and moved to 4.7. Model is tagged. The development evolved as it went along. The datase…
456 likes1,070 downloads in the last 30 daysText generationApache 2.0
This is a synthetic dataset designed for supervised finetuning (SFT) of LLMs. It was used to build SmolLM2-Instruct family of models and contains 1M samples. More detail…
456 likes46,924 downloads in the last 30 daysNo license given
This dataset is a subset of the Open Assistant dataset, which you can find here: This subset of the data only contains the highest-rated paths in the conversation tree,…
452 likes8,471 downloads in the last 30 daysNo license given
纯文本数据，中文医疗数据集，包含预训练数据的百科数据，指令微调数据和奖励模型数据。
449 likes2,376 downloads in the last 30 daysText generationApache 2.0
This dataset is an attempt to replicate the results of Microsoft's Orca Our dataset consists of: ~1 million of FLANv2 augmented with GPT-4 completions (flan1m-alpaca-unc…
444 likes1,651 downloads in the last 30 daysText generationApache 2.0
UltraFeedback is a large-scale, fine-grained, diverse preference dataset, used for training powerful reward models and critic models. We collect about 64k prompts from d…
443 likes12,169 downloads in the last 30 daysText generationMIT
This dataset comprises high-quality audio recordings of Indian speakers reading a standardized 50-word paragraph in four distinct emotional tones — happy, sad, surprised…
437 likes163 downloads in the last 30 daysAudio classificationCC BY 4.0
E-Commerce Customer Support English Audio Dataset Text spoken by all participants: "My order hasn’t arrived yet; can you track it? It was supposed to be delivered yester…
437 likes124 downloads in the last 30 daysAudio classificationCC BY 4.0
Full paper - ICLR 2026 oral Common Corpus is the largest open licensed text dataset, comprising 2.27 trillion tokens (2,267,302,720,836 tokens). It is a diverse dataset,…
436 likes131,361 downloads in the last 30 daysNo license given
The GitHub Code dataest consists of 115M code files from GitHub in 32 programming languages with 60 extensions totalling in 1TB of text data. The dataset was created fro…
433 likes29,697 downloads in the last 30 daysText generationOther license
Airline Customer Support Hinglish Audio Dataset *This dataset contains intentionally low-quality (“B-grade”) data. It has been curated to include noisy, imperfect, or ot…
433 likes73 downloads in the last 30 daysAudio classificationCC BY 4.0
This dataset is focused on improving LLM logical reasoning skills and was used to train the Platypus2 models. It is comprised of the following datasets, which were filte…
428 likes13,125 downloads in the last 30 daysNo license given
FDE-Gym: The Last Mile of AI Delivery FDE-Gym tests whether an agent can do the work of a forward-deployed engineer: enter a customer's running environment, find out wha…
427 likes203 downloads in the last 30 daysFde gym cases termsGated
A new dataset of 7,787 genuine grade-school level, multiple-choice science questions, assembled to encourage research in advanced question-answering. The dataset is part…
424 likes601,170 downloads in the last 30 daysQuestion answeringCC BY-SA 4.0
RedPajama V2: an Open Dataset for Training Large Language Models
420 likes10,532 downloads in the last 30 daysText generationNo license given
Open dataset of 21 Million detailed personas for diverse and controllable synthetic text generation. FinePersonas contains detailed personas for creating customized, rea…
418 likes827 downloads in the last 30 daysText generationLlama 3 license
OctoCodingBench: Instruction-Following Benchmark for Coding Agents English | 中文 🌟 Overview OctoCodingBench benchmarks scaffold-aware instruction following in repository…
417 likes473 downloads in the last 30 daysText generationMIT
UltraData-SFT-2605 is the full set of core-domain SFT data used in the post-training of MiniCPM5-1B-SFT within the MiniCPM5-1B series, and a key representative of L3 ref…
416 likes17,721 downloads in the last 30 daysText generationApache 2.0Gated
SkyPile-150B is a comprehensive, large-scale Chinese dataset specifically designed for the pre-training of large language models. It is derived from a broad array of pub…
415 likes14,915 downloads in the last 30 daysText generationNo license given
Initial release of the Stack. Included 30 programming languages and 18 permissive licenses. Note: Three included licenses (MPL/EPL/LGPL) are considered weak copyleft lic…
413 likes14,303 downloads in the last 30 daysText generationOther licenseGated
This is a reasoning dataset created using Claude Opus 4.5 with a reasoning depth set to high. Some of these questions are from reedmayhew and the rest were generated. Th…
412 likes525 downloads in the last 30 daysNo license given
The HumanEval dataset released by OpenAI includes 164 programming problems with a function sig- nature, docstring, body, and several unit tests. They were handwritten to…
410 likes241,321 downloads in the last 30 daysMIT
WebSight is a large synthetic dataset containing HTML/CSS codes representing synthetically generated English websites, each accompanied by a corresponding screenshot. Th…
406 likes22,130 downloads in the last 30 daysCC BY 4.0
YouTube-Commons is a collection of audio transcripts of 2,063,066 videos shared on YouTube under a CC-By license. Content The collection comprises 22,709,724 original an…
403 likes7,125 downloads in the last 30 daysText generationCC BY 4.0
This is a high-fidelity reasoning dataset synthesized using Claude Opus 4.6. The dataset is designed to capture the model's internal "Chain of Thought" and reasoning tra…
401 likes341 downloads in the last 30 daysMIT
Pre-parsed English and French Wikipedia articles, extracted using the Wikimedia Enterprise Snapshot API. This dataset contains all articles of the English and French lan…
401 likes10,843 downloads in the last 30 daysCC BY-SA 4.0
SWE-bench Verified is a subset of 500 samples from the SWE-bench test set, which have been human-validated for quality. SWE-bench is a dataset that tests systems’ abilit…
400 likes181,895 downloads in the last 30 daysNo license given
A tool for conducting AI-powered qualitative research interviews at scale. In this study, we used Anthropic Interviewer to explore how 1,250 professionals integrate AI i…
399 likes1,331 downloads in the last 30 daysMIT
It's worth noting that our team was working with Glint-Research to collect as much fable data as possible. These are just the anonymized raw traces of both of our teams…
396 likes3,788 downloads in the last 30 daysText generationMIT
This repository contains the dataset used for fine-tuning a language model using SFT for Chain-of-Thought Activation from the paper Generative AI Act II: Test Time Scali…
396 likes504 downloads in the last 30 daysText generationApache 2.0
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 informat…
391 likes149,150 downloads in the last 30 daysText generationODC-By
What is it? 📐 FineMath consists of 34B tokens (FineMath-3+) and 54B tokens (FineMath-3+ with InfiMM-WebMath-3+) of mathematical educational content filtered from Common…
391 likes45,575 downloads in the last 30 daysODC-By
Multi-turn tool-calling trajectories for training AI agents using the Hermes Agent harness. Each sample is a real agent conversation with step-by-step reasoning ( think…
390 likes1,212 downloads in the last 30 daysText generationApache 2.0
May 8, 2025: We found that 48 samples contained hints that revealed the answers. The relevant questions have now been revised to remove the leaked answers. April 14, 202…
389 likes15,721 downloads in the last 30 daysText generationMIT
How to download and use FineVideo Using datasets Using huggingface_hub Load a subset of the dataset Dataset Structure Data Instances Data Fields Dataset Creation License…
386 likes18,714 downloads in the last 30 daysVisual question answeringCCGated
Amazon Review 2023 is an updated version of the Amazon Review 2018 dataset. This dataset mainly includes reviews (ratings, text) and item metadata (desc- riptions, categ…
383 likes79,176 downloads in the last 30 daysNo license given
The Aya Dataset is a multilingual instruction fine-tuning dataset curated by an open-science community via Aya Annotation Platform from Cohere Labs. The dataset contains…
378 likes22,430 downloads in the last 30 daysOtherApache 2.0
XYZ-Aquila SFT is a bilingual release of 7,000 multi-turn, search-oriented tool-use trajectories, comprising 5,000 English examples and 2,000 Chinese examples. This rele…
375 likes592 downloads in the last 30 daysText generationApache 2.0
Arxiver consists of 63,357 arXiv papers converted to multi-markdown (.mmd) format. Our dataset includes original arXiv article IDs, titles, abstracts, authors, publicati…
372 likes2,385 downloads in the last 30 daysCC BY-NC-SA 4.0
This dataset is a collection of approximately 52,00090,000 conversations scraped via the ShareGPT API before it was shut down. These conversations include both user prom…
372 likes7,603 downloads in the last 30 daysText generationCC0 1.0
Books are a rich source of both fine-grained information, how a character, an object or a scene looks like, as well as high-level semantics, what someone is thinking, fe…
370 likes14,371 downloads in the last 30 daysText generationUnknown license
Keiran Paster*, Marco Dos Santos*, Zhangir Azerbayev, Jimmy Ba GitHub | ArXiv | PDF OpenWebMath is a dataset containing the majority of the high-quality, mathematical te…
369 likes36,379 downloads in the last 30 daysNo license given
A compound AI approach to personas grounded in real-world distributions v1.1 Update The v1.1 update introduces the following changes: leverage openai/gpt-oss-120b model…
367 likes4,083 downloads in the last 30 daysText generationCC BY 4.0
The dataset contains problem descriptions and code in python language. This dataset is taken from sahil2801/code_instructions_120k, which adds a prompt column in alpaca…
360 likes19,834 downloads in the last 30 daysQuestion answeringNo license given
🤗 Dataset | 💻 Source Code | 🇨🇳 中文 README UltraData-Math is a large-scale, high-quality mathematical pre-training dataset totaling 290B+ tokens across three progressi…
360 likes29,602 downloads in the last 30 daysText generationApache 2.0
Egocentric-10K is the largest egocentric dataset. It is the first dataset collected exclusively in real factories. Your browser does not support the video tag. Egocentri…
359 likes43,290 downloads in the last 30 daysApache 2.0Gated
The CNN / DailyMail Dataset is an English-language dataset containing just over 300k unique news articles as written by journalists at CNN and the Daily Mail. The curren…
358 likes73,435 downloads in the last 30 daysSummarizationApache 2.0
Our dataset ranked #1 on the Hugging Face Datasets Trending leaderboard from January 20 to January 30. 🚀 Overview The Superior-Reasoning-SFT-gpt-oss-120b dataset is a h…
358 likes1,562 downloads in the last 30 daysText generationCC BY 4.0
We replicated and improved the Berkeley Sky-T1 data pipeline using SFT distillation data from DeepSeek-R1 to create Bespoke-Stratos-17k -- a reasoning dataset of questio…
358 likes26,044 downloads in the last 30 daysApache 2.0
This is a pre-processed version of the UltraFeedback dataset and was used to train Zephyr-7Β-β, a state of the art chat model at the 7B parameter scale. The original Ult…
358 likes20,417 downloads in the last 30 daysText generationMIT
The CT-RATE Team organizes the VLM3D Challenge VLM3D 2026 (2nd Edition) → Challenge Finals at MICCAI 2026 VLM3D 2025 (1st Edition) → Challenge Finals at MICCAI 2025 • Wo…
357 likes110,150 downloads in the last 30 daysImage to textCC BY-NC-SA 4.0Gated
Step-3.5-Flash-SFT is a general-domain supervised fine-tuning release for chat models. This repository keeps the full training interface in one place: json/: canonical r…
357 likes3,350 downloads in the last 30 daysText generationApache 2.0
This dataset contains 10,000 agentic coding and reasoning multi-turn high-quality traces generated by the new Fable 5.1 model using max reasoning effort. It holds almost…
356 likes5,597 downloads in the last 30 daysText generationApache 2.0
The task of PubMedQA is to answer research questions with yes/no/maybe (e.g.: Do preoperative statins reduce atrial fibrillation after coronary artery bypass grafting?)…
351 likes42,031 downloads in the last 30 daysQuestion answeringMIT
本数据应用于项目：Firefly（流萤）: 中文对话式大语言模型 ，训练后得到的模型firefly-1b4 如果您觉得此数据集对您有帮助，请like此数据集并在Github项目中star我们。 我们收集了23个常见的中文数据集，对于每个任务，由人工书写若干种指令模板，保证数据的高质量与丰富度，数据量为115万 。数据分布如下图所示： 每…
351 likes4,139 downloads in the last 30 daysNo license given
This dataset contains a subset of 500 problems from the MATH benchmark that OpenAI created in their Let's Verify Step by Step paper. See their GitHub repo for the source…
350 likes183,357 downloads in the last 30 daysText generationNo license given
Every Hacker News item since 2006, live-updated every 5 minutes What is it? This dataset contains the complete Hacker News archive: every story, comment, Ask HN, Show HN…
350 likes200,679 downloads in the last 30 daysText generationODC-By
CodeSearchNet corpus is a dataset of 2 milllion (comment, code) pairs from opensource libraries hosted on GitHub. It contains code and documentation for several programm…
346 likes42,885 downloads in the last 30 daysText generationOther license
HotpotQA is a new dataset with 113k Wikipedia-based question-answer pairs with four key features: (1) the questions require finding and reasoning over multiple supportin…
346 likes101,970 downloads in the last 30 daysQuestion answeringCC BY-SA 4.0
Mixture-of-Thoughts is a curated dataset of 350k verified reasoning traces distilled from DeepSeek-R1. The dataset spans tasks in mathematics, coding, and science, and i…
345 likes17,095 downloads in the last 30 daysText generationNo license given
Ultra-FineWeb-L3 is the L3 refined data for general high-quality web data within UltraData's L0-L4 tiered data management framework. Moving beyond L2 quality selection,…
342 likes30,941 downloads in the last 30 daysText generationApache 2.0
MMMU (A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI) 🌐 Homepage | 🏆 Leaderboard | 🤗 Dataset | 🤗 Paper | 📖 arXiv | GitHu…
340 likes78,384 downloads in the last 30 daysQuestion answeringApache 2.0
This is an updated and better extracted version of the wikimedia/Wikipedia dataset originally released in 2023. We carefully parsed Wikipedia HTML dumps from August of 2…
336 likes12,000 downloads in the last 30 daysText generationCC BY-SA 4.0
This dataset has been automatically cleaned to remove: Empty or missing responses Responses shorter than 10 characters Refusal responses ("problem is incomplete", "canno…
335 likes788 downloads in the last 30 daysApache 2.0
This dataset contains English Instruction-Following generated by GPT-4 using Alpaca prompts for fine-tuning LLMs. The dataset was originaly shared in this repository: Th…
334 likes3,103 downloads in the last 30 daysText generationCC BY-NC 4.0
弱智吧精选问题数据来自github提供的疑问句，调用GPT-4获取答案，并过滤掉明显拒答的回复。
333 likes215 downloads in the last 30 daysText generationApache 2.0
The dataset contains 12k examples from Orca style dataset Open-Orca/OpenOrca.
331 likes2,222 downloads in the last 30 daysApache 2.0
🚢 Stanford Human Preferences Dataset (SHP) If you mention this dataset in a paper, please cite the paper: Understanding Dataset Difficulty with V-Usable Information (IC…
330 likes13,758 downloads in the last 30 daysText generationNo license given
DCLM-baseline is a 4T token / 3B document pretraining dataset that achieves strong performance on language model benchmarks. Below are comparisions of model trained on D…
328 likes440,863 downloads in the last 30 daysCC BY 4.0
Opus-5.5 generated Doctor-Patient Conversations for All Human Diseases The sequel to nisten/opus-doctor-patient-conversations-all-human-diseases (Opus 4.8). Same disease…
328 likes4,374 downloads in the last 30 daysQuestion answeringApache 2.0
🔉 𝗦𝗟𝗔𝗠 𝗹𝗮𝗯 - 𝗥𝟭-𝗗𝗶𝘀𝘁𝗶𝗹𝗹-𝗦𝗙𝗧 Dataset Lewis Tunstall, Ed Beeching, Loubna Ben Allal, Clem Delangue 🤗 and others at Hugging Face announced today that t…
326 likes4,386 downloads in the last 30 daysCC BY-NC-SA 4.0
[2024/5/8] We released ShareGPT4Video, a large-scale video-caption dataset, with 40K captions annotated by GPT4V and 4.8M captions annotated by our ShareCaptioner-Video.…
323 likes1,981 downloads in the last 30 daysVisual question answeringCC BY-NC 4.0
322 likes715 downloads in the last 30 daysApache 2.0
C-Eval is a comprehensive Chinese evaluation suite for foundation models. It consists of 13948 multi-choice questions spanning 52 diverse disciplines and four difficulty…
321 likes46,614 downloads in the last 30 daysText classificationCC BY-NC-SA 4.0
This is the Open Instruction Generalist Dataset This is our attempt to create a large instruction dataset of medium quality along with a smaller high quality instruciton…
317 likes3,770 downloads in the last 30 daysApache 2.0
UltraData-SFT-Agent-2609 is the L3 refined data for Agent instruction-tuning within UltraData's L0-L4 tiered data management framework. Built for the post-training of Mi…
317 likes20,988 downloads in the last 30 daysText generationApache 2.0
Docmatix is part of the Idefics3 release (stay tuned). It is a massive dataset for Document Visual Question Answering that was used for the fine-tuning of the vision-lan…
316 likes21,276 downloads in the last 30 daysVisual question answeringMIT
🦣 MAmmoTH: Building Math Generalist Models through Hybrid Instruction Tuning MathInstruct is a meticulously curated instruction tuning dataset that is lightweight yet g…
316 likes12,385 downloads in the last 30 daysText generationMIT
We have received a DMCA takedown notice from The Pokémon Company International, Inc. In response to this action, we have taken down the dataset. We appreciate your under…
314 likes9,151 downloads in the last 30 daysText to imageCC BY-NC-SA 4.0Gated
This dataset is designed for ORPO or DPO training. See Fine-tune Llama 3 with ORPO for more information about how to use it. It is a combination of the following high-qu…
314 likes752 downloads in the last 30 daysText generationApache 2.0
Open Assistant Conversations Dataset Release 2 (OASST2) Dataset Structure This dataset contains message trees. Each message tree has an initial prompt message as the roo…
311 likes20,027 downloads in the last 30 daysApache 2.0
Research Plan Generation dataset with three subsets covering ML, Arxiv, and PubMed research papers. Each subset contains research tasks with evaluation rubrics and refer…
310 likes182 downloads in the last 30 daysNo license given
olmOCR-bench is a dataset of 1,403 PDF files, plus 7,010 unit test cases that capture properties of the output that a good OCR system should have. This benchmark evaluat…
309 likes34,698 downloads in the last 30 daysODC-By
An Apache-2.0 dataset curated by Eric Hartford and Cognitive Computations Discord: Sponsors Our appreciation for the generous sponsors of Dolphin R1 - Without whom this…
309 likes757 downloads in the last 30 daysApache 2.0
This is a performance art project. Anthropic built their models on the world's freely shared information, then introduced increasingly dystopian data policies to stop an…
308 likes400 downloads in the last 30 daysText generationMIT
This is a new curated subset of our OpenOrca data. This release provides an efficient means of reaching performance on-par with using larger slices of our data, while on…
307 likes3,438 downloads in the last 30 daysText classificationMIT
This synthetic dataset is a collection of 1.6 millions short and clear code snippets that can help LLM models learn how to reason with both natural and programming langu…
306 likes836 downloads in the last 30 daysText generationMITGated
Developed by: Banghua Zhu * , Evan Frick * , Tianhao Wu * , Hanlin Zhu and Jiantao Jiao. License: Apache-2.0 license under the condition that the dataset is not used to…
305 likes2,214 downloads in the last 30 daysApache 2.0
The world's knowledge in 1+1T tokens of parallel text What is it? This dataset contains over 1 trillion tokens of parallel text in English and 500+ languages. It was obt…
305 likes25,699 downloads in the last 30 daysText generationODC-By
2017–2026 · Quotes, orders and trades · Parquet A historical archive of Chinese exchange Level-2 data, supplied through a vendor export. It includes ten-level quote snap…
305 likes140,039 downloads in the last 30 daysResearch use onlyGated
SYNTH is the first open generalist synthetic dataset for training small reasoning model end-to-end, jointly released by Pleias and the AI Alliance. SYNTH includes 79,648…
302 likes11,403 downloads in the last 30 daysText generationCC BY 4.0
BONES-SEED: Skeletal Everyday Embodiment Dataset BONES-SEED is an open dataset of 142,220 annotated human motion animations for humanoid robotics. It provides motion cap…
301 likes6,379 downloads in the last 30 daysRoboticsBones seed licenseGated
Institutional Books is a growing corpus of public domain books. This release is comprised of 983,004 public domain books digitized as part of Harvard Library's participa…
301 likes6,793 downloads in the last 30 daysNo license givenGated
Help keep USASI useful.
Optional. No USASI account required. Payment takes place on the linked provider’s website (Buy Me a Coffee).
About supporting this project