datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
IFEval
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.MegaMath
MegaMath: Pushing the Limits of Open Math Copora
Megamath is part of TxT360, curated by LLM360 Team.
We introduce MegaMath, an open math pretraining dataset curated from diverse, math-focused sources, with over 300B tokens.
MegaMath is curated via the following three efforts:
Revisiting web data:
We re-extracted mathematical documents from Common Crawl with math-oriented HTML optimizations, fasttext-based filtering and deduplication, all for acquiring higher-quality data on… See the full description on the dataset page: https://huggingface.co/datasets/IFM/MegaMath.Code-Reasoning
Code-Reasoning
Dataset Description
Code problem-solving data with reasoning, direct-answer, and task-synthesis subsets. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards, which supports Dataset Viewer inspection and streaming access.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets
IFM/TxT360-v2
Web and… See the full description on the dataset page: https://huggingface.co/datasets/IFM/Code-Reasoning.Math-Reasoning
Math-Reasoning
Dataset Description
Mathematical problem-solving, rewriting, and dialogue data for reasoning-oriented language-model training. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards, which supports Dataset Viewer inspection and streaming access.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets
IFM/TxT360-v2… See the full description on the dataset page: https://huggingface.co/datasets/IFM/Math-Reasoning.Pretrain-Behaviors
Pretrain-Behaviors
Dataset Description
Behavior-focused text covering reasoning, planning, data science, games, general content, and format rewriting. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards, which supports Dataset Viewer inspection and streaming access.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets… See the full description on the dataset page: https://huggingface.co/datasets/IFM/Pretrain-Behaviors.TxT360-v2
TxT360-v2
Dataset Description
Pre-training sources for the K2 Horizon training data release. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets
IFM/TxT360-v2
Web and question-answering text
3
IFM/Code-Reasoning
Code reasoning and task synthesis
7
IFM/Math-Reasoning… See the full description on the dataset page: https://huggingface.co/datasets/IFM/TxT360-v2.IFBench_test
License
This dataset is licensed under ODC-BY-1.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. This dataset includes output data generated from third party models that are subject to separate terms governing their use.
Citation
Please cite:
@misc{pyatkin2025generalizing,
title={Generalizing Verifiable Instruction Following},
author={Valentina Pyatkin and Saumya Malik and Victoria Graf and Hamish Ivison and… See the full description on the dataset page: https://huggingface.co/datasets/allenai/IFBench_test.K2Datasets
K2 Dataset Card
The following data mix was used to train K2 and achieve results in line with Llama 2 70B.
Dataset Details
K2 was trained on 1.4T tokens across two stages. The data sources and data mix for each stage are listed below.
Dataset Description: Stage 1
Dataset
Starting Tokens
Multiplier
Total Tokens
% of Total
dm-math
4.33B
3x
13B
1%
pubmed-abstracts (from the Pile)
4.77B
3x
14.3B
1.1%
uspto (from the Pile)
4.77B
3x… See the full description on the dataset page: https://huggingface.co/datasets/IFM/K2Datasets.TxT360-Midas
TxT360-MidAS: Mid-training Actual and Synthetic data
Dataset Summary
TxT360-Midas is a mid-training dataset designed to extend language model context length up to 512k tokens while injecting strong reasoning capabilities via synthetic data.
TxT360-Midas was used to mid-train the K2-V2 LLM, yielding base model with strong long-context performance and reasoning abilities. Resulting model demonstrates strong performance on complex mathematical and logic puzzle tasks.… See the full description on the dataset page: https://huggingface.co/datasets/IFM/TxT360-Midas.ifeval-like-data
IFEval Like Data
This dataset contains instruction-response pairs synthetically generated using Qwen/Qwen2.5-72B-Instruct following the style of google/IFEval dataset and verified for correctness with lm-evaluation-harness. The dataset contains two subsets:
default: which contains 550k unfiltered rows synthetically generated with Qwen2.5-72B-Instruct, a few system prompts and MagPie prompting technique. The prompts can contain conflicting instructions as defined in… See the full description on the dataset page: https://huggingface.co/datasets/argilla/ifeval-like-data.AmberDatasets
Amber-Data
This dataset contains the fully prepared data sequence used to train Amber, an
LLM360 model.
About LLM360
LLM360 is an initiative for comprehensive and fully open-sourced LLMs,
where all training details, model checkpoints, intermediate results, and
additional analyses are made available to the community. Our goal is to advance
the field by inviting the community to deepen the understanding of LLMs
together. As the first step of the project… See the full description on the dataset page: https://huggingface.co/datasets/IFM/AmberDatasets.nfcorpusTxT360-3efforts
TxT360-3efforts
Dataset Summary
TxT360-3efforts is a supervised fine-tuning (SFT) dataset designed to train language models with three reasoning efforts (low, medium, high) controllable via chat template. The dataset consists of approximately 10 million documents with 10 billion loss tokens.
All question sources are either collected from permissively licensed public datasets or synthetically generated, and are subsequently quality-filtered, deduplicated, and… See the full description on the dataset page: https://huggingface.co/datasets/IFM/TxT360-3efforts.IF_multi_constraints_upto5
Dataset
This is the IF-RLVR training data, with up to 5 constraints per instruction. The constraints are sampled from IFEval (25) and IFBench-Train (29).
License
This dataset is licensed under ODC-BY-1.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines. This dataset includes output data generated from third party models that are subject to separate terms governing their use.
Citation
Please cite:… See the full description on the dataset page: https://huggingface.co/datasets/allenai/IF_multi_constraints_upto5.FineVisionConcatShuffleIFXSFT-Reasoning
SFT-Reasoning
Dataset Description
Instruction-following and reasoning data prepared for supervised fine-tuning. This repository is part of the K2 Horizon collection.
The repository is organized into multiple subsets. Every subset has a train split backed by Parquet shards, which supports Dataset Viewer inspection and streaming access.
K2 Horizon Dataset Series
Dataset repository
Focus
Subsets
IFM/TxT360-v2
Web and question-answering text… See the full description on the dataset page: https://huggingface.co/datasets/IFM/SFT-Reasoning.IfEvalCode-testsetIFEval-Ko
IFEval-Ko: Korean Instruction-Following Benchmark for LLMs
This dataset is originated from IFEval Dataset
Korean Version README
IFEval-Ko is a Korean adaptation of Google's open-source IFEval benchmark utilized with lm-evaluation-harness framework.
It enables evaluation of large language models (LLMs) for their instruction-following capabilities in the Korean language.
Dataset Details
Original Source: google/IFEvalAdaptation Author: Allganize Inc. LLM TEAM |… See the full description on the dataset page: https://huggingface.co/datasets/allganize/IFEval-Ko.ifc-bench
IFC-Bench
A benchmark dataset for evaluating BIM (Building Information Modeling) comprehension and reasoning capabilities in AI systems. Provides curated IFC models with question-answer pairs across 4 complexity categories for testing BIM-related AI implementations.
Dataset snapshot:
question
ground_truth
ifc_model
project
category
0
What modelling program and IFC standard were used to create this model?
The model was created using...
arc
4351
1
1
What are the… See the full description on the dataset page: https://huggingface.co/datasets/sylvainHellin/ifc-bench.Instruction-Following-IFEval
SEA-IFEval
SEA-IFEval evaluates a model's ability to adhere to constraints provided in the prompt, for example beginning a response with a specific word/phrase or answering with a certain number of sections. It is based on IFEval and was manually translated by native speakers for Indonesian, Javanese, Sundanese, Thai, Tagalog, and Vietnamese.
Supported Tasks and Leaderboards
SEA-IFEval is designed for evaluating chat or instruction-tuned large language models (LLMs).… See the full description on the dataset page: https://huggingface.co/datasets/aisingapore/Instruction-Following-IFEval.RLVR-IFeval
IF Data - RLVR Formatted
This dataset contains instruction following data formatted for use with open-instruct - specifically reinforcement learning with verifiable rewards.
Prompts with verifiable constraints generated by sampling from the Tulu 2 SFT mixture and randomly adding constraints from IFEval.
Part of the Tulu 3 release, for which you can see models here and datasets here.
Dataset Structure
Each example in the dataset contains the standard instruction-tuning… See the full description on the dataset page: https://huggingface.co/datasets/allenai/RLVR-IFeval.ifm-cleaned-pretrain-30B
IFM Cleaned Pretrain — 30B target
credits to https://huggingface.co/datasets/IFM/Pretrain-Behaviors
Status: complete.
Published: 6,155,901 documents; 30,000,015,784 source-annotated tokens.
Target: 30,000,000,000 source-annotated tokens, approximately equal across all seven categories.
This repository contains text only in Parquet: earlier shards were format-cleaned; subsequent shards contain source text without the cleaner. There are no token-ID arrays or binary token shards.… See the full description on the dataset page: https://huggingface.co/datasets/domofon/ifm-cleaned-pretrain-30B.guru-RL-92k
Revisiting Reinforcement Learning for LLM Reasoning from A Cross-Domain Perspective
Dataset Description
Guru is a curated six-domain dataset for training large language models (LLM) for complex reasoning with reinforcement learning (RL). The dataset contains 91.9K high-quality samples spanning six diverse reasoning-intensive domains, processed through a comprehensive five-stage curation pipeline to ensure both domain diversity and reward verifiability.… See the full description on the dataset page: https://huggingface.co/datasets/IFM/guru-RL-92k.Multi-IF
Dataset Summary
We introduce Multi-IF, a new benchmark designed to assess LLMs' proficiency in following multi-turn and multilingual instructions. Multi-IF, which utilizes a hybrid framework combining LLM and human annotators, expands upon the IFEval by incorporating multi-turn sequences and translating the English prompts into another 7 languages, resulting in a dataset of 4501 multilingual conversations, where each has three turns. Our evaluation of 14 state-of-the-art LLMs on… See the full description on the dataset page: https://huggingface.co/datasets/facebook/Multi-IF.oscar_2023_filteredfrom datasets import load_dataset
ds=load_dataset("if001/oscar_2023_filtered")
ds['train']
---
Dataset({
features: ['text'],
num_rows: 312396
})
oscar 2023をfilterしたものhttps://huggingface.co/datasets/oscar-corpus/OSCAR-2301
詳細はコードを参照https://github.com/if001/HojiChar_OSCAR_sample/tree/0.0.4
ifeval-like-dataVideo-IFBench
Video-IFBench
This release contains the evaluation split used for the Video-IFBench main experiments.
Paper: Video-IFBench: Evaluating Instruction Following of Multimodal LLMs in Video Understanding Scenarios
Project page: https://alexios-hub.github.io/Video-IFBench/
Code: https://github.com/Alexios-hub/Video-IFBench
de_ifevalGerman version of IFEVAL dataset based on jzhang86/corrected_ifeval
GPT-4o is used for translation, every data point is human checked.
guru-RL-92k-extra-info-compressed
Revisiting Reinforcement Learning for LLM Reasoning from A Cross-Domain Perspective
Note for this extra-info-compressed data version!
The dataset provided in this repository is specifically intended for use with the latest release of VeRL (v0.4.0). Since VeRL rl_dataset.py processes datasets as datasets.Dataset, it is essential that the structure of all Parquet files remains fully consistent. This repository is designed to meet that requirement.
In this repo, the… See the full description on the dataset page: https://huggingface.co/datasets/IFM/guru-RL-92k-extra-info-compressed.general-reasoning-ift-pairs
Reasoning-IFT Pairs (General Domain)
This dataset provides the largest set of IFT and Reasoning answers pairs for a set of general domain queries (cf: math-domain).It is based on the Infinity-Instruct dataset, an extensive and high-quality collection of instruction fine-tuning data.
We curated 900k queries from the 7M_core subset of Infinity-Instruct, which covers multiple domains including general knowledge, commonsense Q&A, coding, and math.For each query… See the full description on the dataset page: https://huggingface.co/datasets/Scale-or-Reason/general-reasoning-ift-pairs.