datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
QwQ_InfInstruct_Gen_v0use QwQ 32b preview to generate response to answer the question from Infinity-Instruct gen
QwQ-LongCoT-130KAlso have a look on the second version here => QwQ-LongCoT-2
Figure 1: Just a cute picture generate with [Flux](https://huggingface.co/Shakker-Labs/FLUX.1-dev-LoRA-Logo-Design)
Today, I’m excited to release QwQ-LongCoT-130K, a SFT dataset designed for training O1-like large language models (LLMs). This dataset includes about 130k instances, each with responses generated using QwQ-32B-Preview. The dataset is available under the Apache 2.0 license, so feel free to use it as you like.… See the full description on the dataset page: https://huggingface.co/datasets/amphora/QwQ-LongCoT-130K.smoltalk-chinese-QwQ-Distrill
smoltalk-chinese-QwQ-Distrill [中文] [English]
📖Technical Report
smoltalk-chinese-QwQ-Distrill is a Chinese fine-tuning dataset constructed with reference to the SmolTalk-Chinese dataset. It aims to provide high-quality synthetic reasoning data support for training large language models (LLMs). The dataset consists entirely of synthetic data, comprising over 700,000 entries. It is specifically designed to enhance the performance of Chinese LLMs across various tasks… See the full description on the dataset page: https://huggingface.co/datasets/ChinaunicomSoftware/smoltalk-chinese-QwQ-Distrill.Dataset_of_Russian_thinkingRu
RTD
Описание:Russian Thinking Dataset — это набор данных, предназначенный для обучения и тестирования моделей обработки естественного языка (NLP) на русском языке. Датасет ориентирован на задачи, связанные с генерацией текста, анализом диалогов и решением математических и логических задач.
Основная информация:
Сплит: train
Количество записей: 147.046
Цели:
Обучение моделей пониманию русского языка.
Создание диалоговых систем с естественным взаимодействием.… See the full description on the dataset page: https://huggingface.co/datasets/qwqeqw/Dataset_of_Russian_thinking.QWQ-LongCOT-AIMOQWQ-LongCOT-AIMO is a derived dataset created by processing the amphora/QwQ-LongCoT-130K dataset. It filters the original dataset to focus specifically on question-answering pairs where the final answer is a numerical value between 0 and 999, explicitly marked using the \boxed{...} format within the original chain-of-thought answer.
Dataset Structure
Data Splits
The dataset is split into training, validation, and test sets with an 80/10/10 ratio based on the filtered… See the full description on the dataset page: https://huggingface.co/datasets/Floppanacci/QWQ-LongCOT-AIMO.t3-struct-qwq32b
T3: Transformation of Thinking Traces (QwQ-32B)
This dataset contains T3-transformed thinking traces (Structural Normalization) derived from the QwQ-32B model, as introduced in the paper RAG over Thinking Traces Can Improve Reasoning Tasks.
Retrieval-augmented generation (RAG) is traditionally believed to offer limited benefit for reasoning-intensive tasks like math and code. This work challenges that assumption by showing that retrieving thinking traces—intermediate reasoning… See the full description on the dataset page: https://huggingface.co/datasets/narabzad/t3-struct-qwq32b.OpenMathReasoning-QWQ-17k
OpenMathReasoning-QWQ-17k
This dataset contains 17,000 mathematical reasoning problems from the nvidia/OpenMathReasoning dataset, filtered for:
Chain-of-thought (CoT) reasoning type
Generated by QWQ models
Solutions with ≤16,284 tokens
Format
The dataset is formatted for VERL (Versatile Reinforcement Learning) training with the following fields:
data_source: Source identifier (nvidia/OpenMathReasoning-{model})
prompt: List of messages with role and content (chat… See the full description on the dataset page: https://huggingface.co/datasets/YYF42/OpenMathReasoning-QWQ-17k.EVAL-OT-Qwen2.5-7B-Instruct-QwQ-1k_rows-RL
Column Details
Column
Description
question
The question we want the model to answer
answer
The string answer
task
The name of the task the row belongs to
prompt
The prompt we will feed into the model to solve the question
model_responses
An array of strings that the model generated to answer the prompt (usually size of 4 or 34 depending on the evaluation task)
model_responses__eval_is_correct
An array aligned with model_responses containing booleans: True when… See the full description on the dataset page: https://huggingface.co/datasets/SkillFactory/EVAL-OT-Qwen2.5-7B-Instruct-QwQ-1k_rows-RL.SFT_DATA-openthoughts-1k_rows-baseline-QwQ-AnnotatedYou can train using these datasets with LLaMA-Factory if you add this to your data/datasets.json files.
"example_dataset": {
"hf_hub_url": "SkillFactory/SFT_DATA-openthoughts-1k_rows-baseline-QwQ-Annotated",
"formatting": "sharegpt",
"columns": {
"messages": "conversations"
},
"tags": {
"user_tag": "user",
"assistant_tag": "assistant",
"role_tag": "role",
"content_tag": "content"
},
"subset": "sft_train"
}
t3-traces-qwq32b
T3: Raw Thinking Traces (GPT-OSS-120B)
This dataset contains raw thinking traces—intermediate reasoning trajectories—generated by the GPT-OSS-120B model. These traces were produced during problem-solving attempts across various reasoning-intensive benchmarks.
This work was introduced in the paper: RAG over Thinking Traces Can Improve Reasoning Tasks.
The authors propose that Retrieval-Augmented Generation (RAG) is highly effective for reasoning tasks (like math and code) when… See the full description on the dataset page: https://huggingface.co/datasets/narabzad/t3-traces-qwq32b.OpenMathInstruct-2-QwQ
OpenMathInstruct-2-QwQ
Qwen/QwQ-32B-Preview solutions of 107k augmented_math problems of nvidia/OpenMathInstruct-2.
All solutions are validated and agree with the expected_answer field of OpenMathInstruct-2.
squad-qwq-recall-1k
squad-qwq-recall-1k
This dataset is planned to be used as SFT to create the recall-writer model in flow step 1
Recall Writer Flow
Purpose: To synthetically augment data for training a smaller model to effectively recall new knowledge from documents and apply it in the thinking process.
Step 1:
Distill recall traces from the reasoning model
Ask the reasoning model to recall memories related to the question.
The expectation is that the reasoning model, trained… See the full description on the dataset page: https://huggingface.co/datasets/ping98k/squad-qwq-recall-1k.big-math-ppo-mix-40k-qwq-solutions
Big Math PPO Mix 40k QwQ Solutions
Teacher solutions generated by Qwen/QwQ-32B-Preview for the Big Math PPO Mix
prompts, intended for knowledge distillation into a student policy.
This is the combined dataset, merging:
big-math-ppo-mix-30k prompts → 22,584 correct solutions (75.3% of 30,000)
big-math-ppo-mix-extra-10k prompts → 7,986 correct solutions (79.9% of 10,000)
Total: 30,570 verified-correct solutions (40,000 prompts attempted).
Generation
vLLM sampling… See the full description on the dataset page: https://huggingface.co/datasets/christinakopi/big-math-ppo-mix-40k-qwq-solutions.t3-reflect-qwq32b
T3: Transformation of Thinking Traces
This dataset contains thinking traces (intermediate reasoning trajectories) transformed into structured, retrieval-friendly representations using the T3 method. It is designed to be used as a retrieval corpus for Reasoning-augmented Generation (RAG) to improve performance on math, code, and science tasks.
This specific repository contains the Structural Normalization transformation applied to traces generated by the QwQ-32B model. Structural… See the full description on the dataset page: https://huggingface.co/datasets/narabzad/t3-reflect-qwq32b.EVAL_MATH500-OT-Qwen2.5-7B-Instruct-QwQ-10k_rows-RL
Column Details
Column
Description
question
The question we want the model to answer
answer
The string answer
task
The name of the task the row belongs to
prompt
The prompt we will feed into the model to solve the question
model_responses
An array of strings that the model generated to answer the prompt (usually size of 4 or 34 depending on the evaluation task)
model_responses__eval_is_correct
An array aligned with model_responses containing booleans: True when… See the full description on the dataset page: https://huggingface.co/datasets/SkillFactory/EVAL_MATH500-OT-Qwen2.5-7B-Instruct-QwQ-10k_rows-RL.SFT_DATA-openthoughts-10k_rows-baseline-QwQ-AnnotatedYou can train using these datasets with LLaMA-Factory if you add this to your data/datasets.json files.
"example_dataset": {
"hf_hub_url": "SkillFactory/SFT_DATA-openthoughts-10k_rows-baseline-QwQ-Annotated",
"formatting": "sharegpt",
"columns": {
"messages": "conversations"},
"tags": {
"user_tag": "user",
"assistant_tag": "assistant",
"role_tag": "role",
"content_tag": "content"
},
"subset": "sft_train"
}
S1.1-QwQ
S1.1-QwQ
This dataset is a refined dataset based on s1.1-1k. Specifically, we used QwQ-32B to generate reasoning trajectories for each problem in s1.1-1k dataset.
Based on this dataset, we finetuned our new model based on Qwen2.5-32B-Instruct and DeepSeek-R1-Distill-32B series, we now open-source our best model.
Model: S1.1-QwQ-DS.
Evaluation Results
Category
Model
Model SizeAIME 2024
AIME 2025
MATH500
GPQA
Industrial Models
QwQ
32B
80.0
60.0
97.6
68.2… See the full description on the dataset page: https://huggingface.co/datasets/BitStarWalkin/S1.1-QwQ.EVAL_MATH500-OT-Qwen2.5-7B-Instruct-QwQ-1k_rows-RL
Column Details
Column
Description
question
The question we want the model to answer
answer
The string answer
task
The name of the task the row belongs to
prompt
The prompt we will feed into the model to solve the question
model_responses
An array of strings that the model generated to answer the prompt (usually size of 4 or 34 depending on the evaluation task)
model_responses__eval_is_correct
An array aligned with model_responses containing booleans: True when… See the full description on the dataset page: https://huggingface.co/datasets/SkillFactory/EVAL_MATH500-OT-Qwen2.5-7B-Instruct-QwQ-1k_rows-RL.EVAL-OT-Qwen2.5-7B-Instruct-QwQ-10k_rows-RL
Column Details
Column
Description
question
The question we want the model to answer
answer
The string answer
task
The name of the task the row belongs to
prompt
The prompt we will feed into the model to solve the question
model_responses
An array of strings that the model generated to answer the prompt (usually size of 4 or 34 depending on the evaluation task)
model_responses__eval_is_correct
An array aligned with model_responses containing booleans: True when… See the full description on the dataset page: https://huggingface.co/datasets/SkillFactory/EVAL-OT-Qwen2.5-7B-Instruct-QwQ-10k_rows-RL.