datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Arithmetic-Reasoning
SagheerLab/Arithmetic-Reasoning
A high-quality synthetic arithmetic and elementary mathematics reasoning dataset for training and evaluating small language models - not an "ultimate math" claim, but a clean, verified, tiered reasoning dataset where every answer is programmatically checked.
This dataset was built to train 100M-ish models that benefit disproportionately from clean, unambiguous examples. At 50M examples (45M train / 2.5M val / 2.5M test, ~5GB parquet) it is… See the full description on the dataset page: https://huggingface.co/datasets/SagheerLab/Arithmetic-Reasoning.China-K12-STEM-10K-CoT-Reasoning
K12-STEM-CoT-Chinese
1.54M Chinese K12 STEM problems with chain-of-thought solutions, 48% with diagrams.
The largest structured Chinese math/physics/chemistry reasoning dataset.
This is a curated sample (10,000 problems) of the full 1.54M dataset available via API.
Full Dataset Access
Access the full 1,540,000+ problems via API →
This Sample
Full API
Total problems
10,025
1,540,000+
With CoT solutions
10,025
1,490,000+
With diagrams
6,093
740,000+… See the full description on the dataset page: https://huggingface.co/datasets/lfaviate/China-K12-STEM-10K-CoT-Reasoning.openmath-reasoning-medley
OpenMath Reasoning Curated Dataset
This dataset contains curated math-solution generations for problems from
nvidia/OpenMathReasoning.
Overview
Source dataset: nvidia/OpenMathReasoning
Problems and expected answers: preserved from the source dataset
Solutions: generated during curation runs and stored in generated_solution
Per-example model tracking: stored in generation_model
Statistics
Split
Examples
Previously published
Added this upload… See the full description on the dataset page: https://huggingface.co/datasets/PursuitOfDataScience/openmath-reasoning-medley.context
On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
Charlie Zhang, Graham Neubig,
Xiang Yue
Carnegie Mellon University, Language Technologies Institute
Does Reinforcement Learning Truly Extend Reasoning?
This work explores the discrepancy in views on RL's effectiveness in extending language models' reasoning abilities. Some characterize RL as a capability refiner, while others see it as inducing new compositional skills. This challenge… See the full description on the dataset page: https://huggingface.co/datasets/Interplay-LM-Reasoning/context.composition
On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models
Charlie Zhang, Graham Neubig,
Xiang Yue
Carnegie Mellon University, Language Technologies Institute
Does Reinforcement Learning Truly Extend Reasoning?
This work explores the discrepancy in views on RL's effectiveness in extending language models' reasoning abilities. Some characterize RL as a capability refiner, while others see it as inducing new compositional skills. This challenge… See the full description on the dataset page: https://huggingface.co/datasets/Interplay-LM-Reasoning/composition.olympiad_style_integer_math_reasoning
Olympiad Math Reasoning Traces
Version: v1.0.2
Release date: 2026-04-19
64,763 full model reasoning traces for olympiad-style math problems with verified integer answers. This dataset contains only correct and non-truncated traces — every record contains a terminal \boxed{...} answer (within the last 500 characters of the response) that matches the expected integer exactly, and none of the responses hit the model's generation-token cap. Intended for distillation and supervised… See the full description on the dataset page: https://huggingface.co/datasets/mihailgribov/olympiad_style_integer_math_reasoning.reasoning-lite
Anonymous Reasoning Lite
This repository contains data accompanying an anonymous TMLR submission. It provides
1,211,520 sampled reasoning attempts from four model configurations on competition-math
and field-balanced multiple-choice questions. The data omits top-20 alternative-token
distributions while retaining realized-token log probabilities and ranks.
Contents
The same attempts are available in full and metadata-only representations:
Configuration group… See the full description on the dataset page: https://huggingface.co/datasets/AnonymizedTMLRSubmission/reasoning-lite.2d_3d_seq_path_spatial_reasoning
Spatial Reasoning Dataset
A synthetic dataset of Hamiltonian path puzzles with rich chain-of-thought reasoning, designed for training and evaluating spatial reasoning in language models.
Overview
Each sample presents a grid-based puzzle where the solver must find a path visiting every cell exactly once, moving only up/down/left/right (plus above/below for 3D). Puzzles span 2D grids (3x3 to 8x8) and 3D cubes (3x3x3 to 4x4x4), covering solvable, impossible, and multi-turn… See the full description on the dataset page: https://huggingface.co/datasets/eousphoros/2d_3d_seq_path_spatial_reasoning.verified-research-reasoning-trajectories
Verified Research Reasoning Trajectories for RLVR
This repository is the public sample and schema repository for Ulam's research-level mathematical reasoning trajectories for reinforcement learning with verifiable rewards (RLVR), process supervision, judge training, proof criticism, and private evaluations.
Ulam Verified Research Reasoning Trajectories are proof-process data for RLVR. Each record contains a normalized research problem, a golden or partial-golden proof graph… See the full description on the dataset page: https://huggingface.co/datasets/ulamai/verified-research-reasoning-trajectories.ReasonMapThis repository contains the dataset described in the paper Can MLLMs Guide Me Home? A Benchmark Study on Fine-Grained Visual
Reasoning from Transit Maps.
Project page: https://fscdc.github.io/Reason-Map
Code: https://github.com/fscdc/ReasonMap
Please refer json file in GitHub Repo under stations path.
Deepseek-V4-Reasoning-Code-2500
DeepSeek Reasoning and Code Distillation Dataset
This dataset contains synthetic instruction-response examples generated from coding, reasoning, and math prompts. It was generated with enforce_distillable_text enabled using DeepSeek V4 Pro and DeepSeek V4 Flash through OpenRouter. It is intended for experimentation with supervised fine-tuning, response-style distillation, reasoning-format analysis, and code-assistant behavior research.
The dataset file is:
train.csv
It contains 2… See the full description on the dataset page: https://huggingface.co/datasets/Banaxi-Tech/Deepseek-V4-Reasoning-Code-2500.chart-reasoning-verified
chart-reasoning-verified
Chart reasoning examples generated from an explicit latent representation.
The data, the question and the answer are computed before the chart is
drawn, so the image is a rendering of known ground truth rather than the
source of it. No model was asked to label anything.
Each row carries both a rendered chart and a text serialisation of the same
chart, so the set is usable for vision-language training and for text-only
language model training without… See the full description on the dataset page: https://huggingface.co/datasets/vinod-anbalagan/chart-reasoning-verified.tool-reasoning-sft-TOOLS-toolace-sft-tool-use-agent-data-cleaned-rectified
ToolACE - Tool-Use Agent Data Cleaned & Rectified
👥 Follow the Author
Aman Priyanshu
Overview
This dataset is a cleaned and restructured version of the Team-ACE/ToolACE dataset. ToolACE is a high-quality conversational tool-use dataset containing 11,300+ examples of natural language interactions requiring function calling across diverse domains. This version converts the original OpenAI function-call format into a standardized multi-turn tool-use… See the full description on the dataset page: https://huggingface.co/datasets/AmanPriyanshu/tool-reasoning-sft-TOOLS-toolace-sft-tool-use-agent-data-cleaned-rectified.pentabrid-reproducibility
Pentabrid 27B: reproducibility package
Everything required to recompute the results of a controlled evaluation of fine-tuning
configurations for medical question answering. Openly available with no access
restrictions.
Contents
Path
Description
per_item/medxpertqa_*.jsonl
Per-item predictions for all six checkpoints on 2,450 MedXpertQA-Text items. Fields: id, gold, extracted_answer, correct, explicit_marker_present, n_markers, response_chars… See the full description on the dataset page: https://huggingface.co/datasets/Clinical-Reasoning-Hub/pentabrid-reproducibility.sober_reasoning
🧠 Sober Reasoning: Evaluation Logs
This repository hosts evaluation logs and outputs from our paper:
"A Sober Look at Progress in Language Model Reasoning: Pitfalls and Paths to Reproducibility"
📄 Paper📊 Leaderboard💻 Evaluation Code
🗂️ Repository Structure
Evaluation logs are organized by the cluster used during inference to highlight hardware-induced variance in model performance (see Section 3.3 of the paper).
sober_reasoning/
├── cluster_A/
│ ├──… See the full description on the dataset page: https://huggingface.co/datasets/bethgelab/sober_reasoning.harmonic-reasoning-v1
Harmonic Reasoning v1
Support This Work
I'm a PhD student in visual neuroscience at the University of Toronto who also happens to spend way too much time fine-tuning, merging, and quantizing open-weight models on rented H100s and a local DGX Spark. All training compute is self-funded — balancing GPU costs against a student budget. If my open-weight models or datasets have been useful to you, consider supporting future releases.
Support on Ko-fi
Harmonic Reasoning v1 is a… See the full description on the dataset page: https://huggingface.co/datasets/DJLougen/harmonic-reasoning-v1.LogicMind-Chat-Reasoning-SFT-300K
Nemotron-Post-Training-Dataset-v2-chat Dataset Card
Overview 📌
This dataset contains 296,168 chat-style instruction/response samples generated by qwen-3-32b. Each record provides a user prompt, an explicit reasoning trace, and a final answer, plus precomputed length fields. The data is packaged as JSONL (one JSON object per line).
Highlights
Scale: 296,168 samples
Category: chat (100%)
Generator: qwen-3-32b (100%)
Structure: problem → qwen3-reasoning → qwen3-solution… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/LogicMind-Chat-Reasoning-SFT-300K.Math_CoT_Arabic_English_Reasoning
Math CoT Arabic English Dataset
A high-quality, bilingual (English & Arabic) dataset for Chain-of-Thought (COT) reasoning in mathematics and related disciplines, developed by Miscovery AI.
Overview
Math-COT is a unique dataset designed to facilitate and benchmark the development of chain-of-thought reasoning capabilities in language models across mathematical domains. With meticulously crafted examples, explicit reasoning steps, and bilingual support, this dataset offers… See the full description on the dataset page: https://huggingface.co/datasets/miscovery/Math_CoT_Arabic_English_Reasoning.K-Paths-inductive-reasoning-drugbank
🔗 This dataset is part of the study:
K-Paths: Reasoning over Graph Paths for Drug Repurposing and Drug Interaction Prediction
📖 Read the Paper
💾 GitHub Repository
DrugBank: Inductive Reasoning Dataset
This dataset contains drug pairs annotated with 86 pharmacological relationships (e.g.,DrugA may increase the anticholinergic activities of DrugB).
Each entry includes two drugs, an interaction label, drug descriptions, and structured/natural language representations… See the full description on the dataset page: https://huggingface.co/datasets/Tassy24/K-Paths-inductive-reasoning-drugbank.multilevel-legal-reasoning
Legal Reasoning Dataset with Multilevel Human and Model-Annotated Explanations
Prepared by Mst Rafia Islam, Umong Sain, Azmine Toushik Wasi
Prepared as a part of Reasoning Datasets Competition by Bespoke Labs, Hugging Face, and Together.ai.
🧭 Purpose and Scope
The Legal Reasoning Dataset aims to support the evaluation and training of legal reasoning systems, particularly in multilingual or jurisdiction-agnostic contexts. It focuses on international acts and treaties… See the full description on the dataset page: https://huggingface.co/datasets/ciol-research/multilevel-legal-reasoning.tinysynth-reasoning
TinySynth Reasoning Primitives
Synthetic training data for teaching small language models stable state
representation and controlled reasoning operations — entity/attribute
binding, state persistence, mutation, transfer, reference resolution,
current-vs-cumulative distinctions, and claim validation — in a
systems/computing vocabulary.
Every example is generated from a hidden symbolic world and verified by a
symbolic solver before any natural language is produced:
semantic state… See the full description on the dataset page: https://huggingface.co/datasets/projenix/tinysynth-reasoning.turkish-seo-reasoning-benchmark-results
Turkish SEO Reasoning Benchmark Results
Bu dataset, Turkish SEO Reasoning benchmark'ının altı farklı model/checkpoint üzerinde çalıştırılmış ham tahminlerini, metriklerini ve tekrar üretim manifestlerini içerir.
Fine-tuned model: berkbirkan/gemma-3-1b-turkish-seo-reasoning-lora
Sonuç
Fine-tuned Gemma 3 1B modeli 22,23 skorla ilk sırada yer aldı. Aynı base model 11,96 skor elde etti.
Mutlak artış: +10,28 puan
Göreli artış: %85,97
Fine-tuned model hata sayısı:… See the full description on the dataset page: https://huggingface.co/datasets/berkbirkan/turkish-seo-reasoning-benchmark-results.GPT-OSS-20B-Distilled-Reasoning-Mini
Dataset Card for Dataset Name
GPT-OSS-20B Distilled Reasoning Dataset Mini
(Multi-stage Evaluative Refinement Method for Reasoning Generation)
Dataset Details and Description
This is a high-quality instruction fine-tuning dataset constructed through knowledge distillation, featuring detailed Chain-of-Thought (CoT) reasoning processes. The dataset is designed to enhance the capabilities of smaller language models in complex reasoning, logical analysis, and instruction… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/GPT-OSS-20B-Distilled-Reasoning-Mini.Qwill-RP-CreativeWriting-Reasoning
Qwill RP CreativeWriting Reasoning Dataset
📝 Dataset Summary
Qwill-RP-CreativeWriting-Reasoning is a creative writing dataset focused on structured reasoning. Each row contains a fictional or narrative prompt sourced from nothingiisreal/Reddit-Dirty-And-WritingPrompts, along with an AI-generated response that includes:
Reasoning, wrapped in <think>...</think>
Final Answer, wrapped in <answer>...</answer>
The goal is to train or evaluate models on chain-of-thought… See the full description on the dataset page: https://huggingface.co/datasets/marcuscedricridia/Qwill-RP-CreativeWriting-Reasoning.reasoning-spectrum-qa
Reasoning Spectrum QA Dataset
This is a unified dataset of exactly 1,000 examples designed for evaluating reasoning capabilities across different dimensions (factual, commonsense, science, arithmetic, multi-hop, and extractive span).
Dataset Purpose
The dataset brings together questions of varying difficulty and reasoning family types under a single unified schema to facilitate standardized testing and evaluation of large language models.
Source… See the full description on the dataset page: https://huggingface.co/datasets/avreymi/reasoning-spectrum-qa.China-K12-STEM-10K-CoT-Reasoning
K12-STEM-CoT-Chinese
1.54M Chinese K12 STEM problems with chain-of-thought solutions, 48% with diagrams.
The largest structured Chinese math/physics/chemistry reasoning dataset.
This is a curated sample (10,000 problems) of the full 1.54M dataset available via API.
Full Dataset Access
Access the full 1,540,000+ problems via API →
This Sample
Full API
Total problems
10,025
1,540,000+
With CoT solutions
10,025
1,490,000+
With diagrams
6,093
740… See the full description on the dataset page: https://huggingface.co/datasets/a13905873166/China-K12-STEM-10K-CoT-Reasoning.K-Paths-inductive-reasoning-pharmaDB
🔗 This dataset is part of the study:
K-Paths: Reasoning over Graph Paths for Drug Repurposing and Drug Interaction Prediction
📖 Read the Paper
💾 GitHub Repository
PharmacotherapyDB: Inductive Reasoning Dataset
PharmacotherapyDB is a drug repurposing dataset containing drug–disease treatment relations in three categories (disease-modifying, palliates, or non-indication).
Each entry includes a drug and a disease, an interaction label, drug, disease descriptions, and… See the full description on the dataset page: https://huggingface.co/datasets/Tassy24/K-Paths-inductive-reasoning-pharmaDB.gsm8k-SynthLabs-reasoning
GSM8K-SynthLabs
This dataset is a refined version of the GSM8K dataset, enriched with complex reasoning traces in the style of Pleias/SYNTH. It is designed for fine-tuning large language models to improve their reasoning capabilities using a structured, step-by-step thinking process.
Key Features
SYNTH Reasoning: Each problem contains a detailed reasoning trace generated by DeepSeek-V3.2, following the structured format (e.g., Query Parsing, Decomposition… See the full description on the dataset page: https://huggingface.co/datasets/mkurman/gsm8k-SynthLabs-reasoning.samantha-r01-recursive-reasoning-corpus
Samantha R01 Recursive Reasoning Corpus
Answer-only corpus for the first isolated Samantha silent-tick / recursive-latent-reasoning validation. It is normalized for pre_train_recursive_reasoning.py and intentionally contains no visible chain-of-thought or source rationale fields.
The repository is private because it combines sources with mixed or unspecified redistribution terms. Access does not supersede any upstream license.
Split policy
train: 50,000… See the full description on the dataset page: https://huggingface.co/datasets/BRlkl/samantha-r01-recursive-reasoning-corpus.Qwen3-Reasoning-Distill-Q-A-Dataset
Qwen3 Reasoning Distill Q&A Dataset
Repository: RefinedNeuro/Qwen3-Reasoning-Distill-Q-A-Dataset
Authors
Mehmet Can Farsak
Serhat Atayeter
License
This dataset is released under CC0 1.0 Universal (CC0 1.0) Public Domain Dedication.
Dataset Summary
This dataset contains question-answer pairs across six STEM subjects designed for Turkish-language reasoning tasks. It was generated using the qwen3-32b model and is intended for fine-tuning the RN_TR_R2… See the full description on the dataset page: https://huggingface.co/datasets/RefinedNeuro/Qwen3-Reasoning-Distill-Q-A-Dataset.