datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
dclm-baseline-1.0
DCLM-baseline
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 DCLM-baseline with other models in the 7B regime.
Model
Params
Tokens
Open dataset?
CORE
MMLU
EXTENDED
Open weights, closed datasets
Llama2
7B
2T
✗
49.2
45.8
34.1
DeepSeek
7B
2T
✗
50.7
48.5
35.3
Mistral-0.3
7B
?
✗
57.0
62.7
45.1
QWEN-2
7B
?
✗
57.5
71.9
50.5
Llama3
8B
15T
✗… See the full description on the dataset page: https://huggingface.co/datasets/mlfoundations/dclm-baseline-1.0.Rosetta-Activations
Rosetta Activations
Updated: 2026-06-15 02:30 UTC
Contrastive activation extractions for 17 semantic concepts across 46 language models,
supporting cross-architecture mechanistic interpretability research.
Companion concept pair corpus: jamesrahenry/Rosetta_Concept_Pairs
Papers: forthcoming
Dataset Structure
Rosetta-Activations/
├── rcp_v1/ # Current extraction line — richest data (N≈2000)
│ └── {Model_Name}/
│ ├── calibration_{concept}.npy… See the full description on the dataset page: https://huggingface.co/datasets/james-ra-henry/Rosetta-Activations.cybergymIFEval
Dataset Card for IFEval
Dataset Summary
This dataset contains the prompts used in the Instruction-Following Eval (IFEval) benchmark for large language models. It contains around 500 "verifiable instructions" such as "write in more than 400 words" and "mention the keyword of AI at least 3 times" which can be verified by heuristics. To load the dataset, run:
from datasets import load_dataset
ifeval = load_dataset("google/IFEval")
Supported Tasks and… See the full description on the dataset page: https://huggingface.co/datasets/google/IFEval.PIN-200M
PIN-200M
A mini version of "PIN: A Knowledge-Intensive Dataset for Paired and Interleaved Multimodal Documents"
Paper: https://arxiv.org/abs/2406.13923
This dataset contains around 200M samples in PIN format, with around 312 TB storage.
🚀 News
[ 2025.09.22 ] !NEW! 🔥 We have completed the final version of the PIN-200M dataset and conducted some simple statistics on it.
[ 2024.12.06 ] !NEW! 🔥 We have updated the quality signals, enabling a swift assessment of whether a sample meets… See the full description on the dataset page: https://huggingface.co/datasets/m-a-p/PIN-200M.MATH-500
Dataset Card for MATH-500
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 file: https://github.com/openai/prm800k/tree/main?tab=readme-ov-file#math-splits
iclr-wm-backup-public
ICLR Watermark Benchmark — backup overflow (public part)
Companion to the private repo Aak975/iclr-wm-backup, which reached its
storage quota. Together the two repos form ONE backup — every file exists in
exactly one of them, with the same layout:
archives/<sub>/part-0000 ... part-NNNN, MANIFEST.json
restore one archive: cat part-* | zstd -d | tar -x
MANIFEST.json = {"parts": N, "sha256": <whole-stream>, "total_bytes": M}
This public part holds only shareable image data… See the full description on the dataset page: https://huggingface.co/datasets/Aak975/iclr-wm-backup-public.benchmark-bcplusapex-agents
APEX–Agents
APEX–Agents is a benchmark from Mercor for evaluating whether AI agents can execute long-horizon, cross-application professional services tasks. Tasks were created by investment banking analysts, management consultants, and corporate lawyers, and require agents to navigate realistic work environments with files and tools (e.g., docs, spreadsheets, PDFs, email, chat, calendar).
Tasks: 480 total (160 per job category)
Worlds: 33 total (10 banking, 11 consulting, 12… See the full description on the dataset page: https://huggingface.co/datasets/mercor/apex-agents.Darwin-CCrequests
Open LLM Leaderboard Requests
This repository contains the request files of models that have been submitted to the Open LLM Leaderboard.
You can take a look at the current status of your model by finding its request file in this dataset. If your model failed, feel free to open an issue on the Open LLM Leaderboard! (We don't follow issues in this repository as often)
Evaluation Methodology
The evaluation process involves running your models against several benchmarks from… See the full description on the dataset page: https://huggingface.co/datasets/open-llm-leaderboard-old/requests.catalog
Mesh-LLM Catalog
This dataset is the Hugging Face-backed catalog for Mesh-LLM.
The runtime catalog entries live under entries/**/*.json. The Dataset Viewer
uses catalog_rows.jsonl, a flat generated table with one row per model variant.
The catalog deliberately excludes raw blob URLs. Entries should resolve to
Hugging Face repositories and canonical Mesh refs.
pile-uncopyrighted
Pile Uncopyrighted
In response to authors demanding that LLMs stop using their works, here's a copy of The Pile with all copyrighted content removed.Please consider using this dataset to train your future LLMs, to respect authors and abide by copyright law.Creating an uncopyrighted version of a larger dataset (ie RedPajama) is planned, with no ETA.
MethodologyCleaning was performed by removing everything from the Books3, BookCorpus2, OpenSubtitles, YTSubtitles, and OWT2… See the full description on the dataset page: https://huggingface.co/datasets/monology/pile-uncopyrighted.HelpSteer2
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 become more helpful, factually correct and coherent, while being adjustable in terms of the complexity and verbosity of its responses.
This dataset has been created in partnership with Scale AI.
When used to tune a Llama 3.1 70B Instruct Model, we achieve 94.1% on RewardBench, which makes it the best Reward Model as… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/HelpSteer2.sts12-sts
STS12
An MTEB dataset
Massive Text Embedding Benchmark
SemEval-2012 Task 6.
Task category
t2t
Domains
Encyclopaedic, News, Written
Reference
https://www.aclweb.org/anthology/S12-1051.pdf
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code:
import mteb
task = mteb.get_tasks(["STS12"])
evaluator = mteb.MTEB(task)
model = mteb.get_model(YOUR_MODEL)
evaluator.run(model)
To learn more about how to… See the full description on the dataset page: https://huggingface.co/datasets/mteb/sts12-sts.LongBench-v2
LongBench v2: Towards Deeper Understanding and Reasoning on Realistic Long-context Multitasks
🌐 Project Page: https://longbench2.github.io
💻 Github Repo: https://github.com/THUDM/LongBench
📚 Arxiv Paper: https://arxiv.org/abs/2412.15204
LongBench v2 is designed to assess the ability of LLMs to handle long-context problems requiring deep understanding and reasoning across real-world multitasks. LongBench v2 has the following features: (1) Length: Context length ranging from 8k to… See the full description on the dataset page: https://huggingface.co/datasets/zai-org/LongBench-v2.MetaMathQAView the project page:
https://meta-math.github.io/
see our paper at https://arxiv.org/abs/2309.12284
Note
All MetaMathQA data are augmented from the training sets of GSM8K and MATH.
None of the augmented data is from the testing set.
You can check the original_question in meta-math/MetaMathQA, each item is from the GSM8K or MATH train set.
Model Details
MetaMath-Mistral-7B is fully fine-tuned on the MetaMathQA datasets and based on the powerful Mistral-7B model. It is… See the full description on the dataset page: https://huggingface.co/datasets/meta-math/MetaMathQA.aime25
AIME 25
American Invitational Mathematics Examination (AIME) 2025
Citation
If you use the AIME25 dataset in your research, please consider citing it as follows:
@misc{aime25,
title={American Invitational Mathematics Examination (AIME) 2025},
author={Zhang, Yifan and Math-AI, Team},
year={2025},
}
PresentBench
PresentBench: A Fine-Grained Rubric-Based Benchmark for Slide Generation
[🌐 Homepage] [📖 Paper] [💻 Code]
This repository hosts the PresentBench benchmark dataset.
📄 Abstract
Slides serve as a critical medium for conveying information in presentation-oriented scenarios such as academia, education, and business. Despite their importance, creating high-quality slide decks remains time-consuming and cognitively demanding. Recent advances in generative models, such… See the full description on the dataset page: https://huggingface.co/datasets/PresentBench/PresentBench.TOFU
TOFU: Task of Fictitious Unlearning 🍢
The TOFU dataset serves as a benchmark for evaluating unlearning performance of large language models on realistic tasks. The dataset comprises question-answer pairs based on autobiographies of 200 different authors that do not exist and are completely fictitiously generated by the GPT-4 model. The goal of the task is to unlearn a fine-tuned model on various fractions of the forget set.
Quick Links
Website: The landing page for TOFU… See the full description on the dataset page: https://huggingface.co/datasets/locuslab/TOFU.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.
Quick Start… See the full description on the dataset page: https://huggingface.co/datasets/UnipatAI/RoadmapBench.DS-1000 DS-1000 in simplified format
🔥 Check the leaderboard from Eval-Arena on our project page.
See testing code and more information (also the original fill-in-the-middle/Insertion format) in the DS-1000 repo.
Reformatting credits: Yuhang Lai, Sida Wang
forecast-news
Forecast News
Deduplicated daily news corpus used by forecast-sim and future-sim.
Snapshot
31,859,020 articles
3,463 daily partitions
Coverage: 2016-08-26 through 2026-08-31
Snapshot published: 2026-09-18
Stored data size: approximately 158.5 GiB
The Parquet files are the canonical complete representation. The repository also
contains daily JSONL files where available and compact headline JSON files used
by article-browsing workflows.
Layout
Files… See the full description on the dataset page: https://huggingface.co/datasets/shash42/forecast-news.ABot-World-Explorer-500h
ABot World Explorer 500h
ABot World Explorer 500h contains 30,969 action-conditioned video episodes
associated with the data infrastructure described in
ABot-World-0. Each episode preserves an MP4,
dataset-native keyboard actions, captions, and one COLMAP text sparse model.
Dataset facts
Item
Value
Episodes
30,969
Source objects
185,814
Semantic splits
None
License
Apache-2.0
The repository name is an identifier, not an audited… See the full description on the dataset page: https://huggingface.co/datasets/acvlab/ABot-World-Explorer-500h.deception-probes-activations
Deception Probes Activations
Pre-extracted residual-stream activations for training and evaluating deception
detection probes on LLMs. Each example contains per-token hidden states from a
specific transformer layer, saved in bfloat16 safetensors format.
License
This dataset contains activations derived from multiple sources with different licenses.
See the LICENSE file for full details.
Component
Source
License
Apollo Probe Pairs (statements)
Azaria & Mitchell… See the full description on the dataset page: https://huggingface.co/datasets/xycoord/deception-probes-activations.swe-bench-dummy-test-datasetFineFineWeb-sample
FineFineWeb: A Comprehensive Study on Fine-Grained Domain Web Corpus
arXiv: Coming Soon
Project Page: Coming Soon
Blog: Coming Soon
Data Statistics
Domain (#tokens/#samples)
Iteration 1 Tokens
Iteration 2 Tokens
Iteration 3 Tokens
Total Tokens
Iteration 1 Count
Iteration 2 Count
Iteration 3 Count
Total Count
aerospace
5.77B
261.63M
309.33M
6.34B
9100000
688505
611034
10399539
agronomy
13.08B
947.41M
229.04M
14.26B
15752828
2711790
649404
19114022
artistic… See the full description on the dataset page: https://huggingface.co/datasets/m-a-p/FineFineWeb-sample.hermes-function-calling-v1
Hermes Function-Calling V1
This dataset is the compilation of structured output and function calling data used in the Hermes 2 Pro series of models.
This repository contains a structured output dataset with function-calling conversations, json-mode, agentic json-mode and structured extraction samples, designed to train LLM models in performing function calls and returning structured output based on natural language instructions. The dataset features various conversational scenarios… See the full description on the dataset page: https://huggingface.co/datasets/NousResearch/hermes-function-calling-v1.nfcorpus
NFCorpus
An MTEB dataset
Massive Text Embedding Benchmark
NFCorpus: A Full-Text Learning to Rank Dataset for Medical Information Retrieval
Task category
t2t
Domains
Medical, Academic, Written
Reference
https://www.cl.uni-heidelberg.de/statnlpgroup/nfcorpus/
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code:
import mteb
task = mteb.get_tasks(["NFCorpus"])
evaluator = mteb.MTEB(task)
model =… See the full description on the dataset page: https://huggingface.co/datasets/mteb/nfcorpus.WavCaps
WavCaps
WavCaps is a ChatGPT-assisted weakly-labelled audio captioning dataset for audio-language multimodal research, where the audio clips are sourced from three websites (FreeSound, BBC Sound Effects, and SoundBible) and a sound event detection dataset (AudioSet Strongly-labelled Subset).
Paper: https://arxiv.org/abs/2303.17395
Github: https://github.com/XinhaoMei/WavCaps
Statistics
Data Source
# audio
avg. audio duration (s)avg. text length
FreeSound… See the full description on the dataset page: https://huggingface.co/datasets/cvssp/WavCaps.