Dataset Card for s1K

  Dataset Summary

s1K is a dataset of 1,000 examples of diverse, high-quality & difficult questions with distilled reasoning traces & solutions from Gemini Thining. Refer to the s1 paper for more details.

  Usage

pip install -q datasets

from datasets import load_dataset ds = load_dataset("simplescaling/s1K")["train"] ds[0]

  Dataset Structure





  Data Instances

An example looks as follows: { 'solution': '1. **Rewrite… See the full description on the dataset page: https://huggingface.co/datasets/simplescaling/s1K.

Citation Information

@misc{muennighoff2025s1simpletesttimescaling, title={s1: Simple test-time scaling}, author={Niklas Muennighoff and Zitong Yang and Weijia Shi and Xiang Lisa Li and Li Fei-Fei and Hannaneh Hajishirzi and Luke Zettlemoyer and Percy Liang and Emmanuel Candès and Tatsunori Hashimoto}, year={2025}, eprint={2501.19393}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2501.19393}, }

Summary:

The open-s1 dataset contains 18,615 mathematical reasoning problems, filtered from the s1K dataset. It’s part of the Open RS project, aimed at enhancing reasoning in small LLMs using reinforcement learning.

Original dataset: https://huggingface.co/datasets/knoveleng/open-s1

Citation Information

@misc{muennighoff2025s1simpletesttimescaling, title={s1: Simple test-time scaling}, author={Niklas Muennighoff and Zitong Yang and Weijia Shi and Xiang Lisa Li and Li Fei-Fei and Hannaneh Hajishirzi and Luke Zettlemoyer and Percy Liang and Emmanuel Candès and Tatsunori Hashimoto}, year={2025}, eprint={2501.19393}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2501.19393}, }

Dataset Card for s1K

  Dataset Summary

s1K-1.1 consists of the same 1,000 questions as in s1K but with traces instead generated by DeepSeek r1. We find that these traces lead to much better performance.

  Usage

pip install -q datasets

from datasets import load_dataset ds = load_dataset("simplescaling/s1K-1.1")["train"] ds[0]

  Dataset Structure





  Data Instances

An example looks as follows: { 'solution': '1. **Rewrite the function using… See the full description on the dataset page: https://huggingface.co/datasets/simplescaling/s1K-1.1.

The benchmark constructed in paper S1-Bench: A Simple Benchmark for Evaluating System 1 Thinking Capability of Large Reasoning Models.

  Introduction

S1-Bench is a novel benchmark designed to evaluate Large Reasoning Models' performance on simple tasks that favor intuitive system 1 thinking rather than deliberative system 2 reasoning. S1-Bench comprises 422 question-answer pairs across four major categories and 28 subcategories, balanced with 220 English and 202 Chinese questions.… See the full description on the dataset page: https://huggingface.co/datasets/WYRipple/S1-Bench.

presencesw/en_processed_open-s1 dataset hosted on Hugging Face and contributed by the HF Datasets community

ming0100/yt-s1 dataset hosted on Hugging Face and contributed by the HF Datasets community

Seanie-lee/ThinkSafe-0.6B-s1 dataset hosted on Hugging Face and contributed by the HF Datasets community

yiboowang/R1-Compress-s1 dataset hosted on Hugging Face and contributed by the HF Datasets community

donoway/s1K-s1.1-32B-rollouts dataset hosted on Hugging Face and contributed by the HF Datasets community

seosiju/BigEarthNet-S1 dataset hosted on Hugging Face and contributed by the HF Datasets community

ljvmiranda921/msde-S1-cs dataset hosted on Hugging Face and contributed by the HF Datasets community

ydeng9/s1-rd dataset hosted on Hugging Face and contributed by the HF Datasets community

mlfoundations-dev/s1-1.1-literally-1-example dataset hosted on Hugging Face and contributed by the HF Datasets community

Citation Information

@misc{muennighoff2025s1simpletesttimescaling, title={s1: Simple test-time scaling}, author={Niklas Muennighoff and Zitong Yang and Weijia Shi and Xiang Lisa Li and Li Fei-Fei and Hannaneh Hajishirzi and Luke Zettlemoyer and Percy Liang and Emmanuel Candès and Tatsunori Hashimoto}, year={2025}, eprint={2501.19393}, archivePrefix={arXiv}, primaryClass={cs.CL}, url={https://arxiv.org/abs/2501.19393}, }

jdp22/s1-tool-new dataset hosted on Hugging Face and contributed by the HF Datasets community

simplescaling/s1K-1.1_tokenized dataset hosted on Hugging Face and contributed by the HF Datasets community

zzzeeee/s1-tool-1k-latest dataset hosted on Hugging Face and contributed by the HF Datasets community

siyanzhao/s1-59k-minus-s1k dataset hosted on Hugging Face and contributed by the HF Datasets community