datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
DeepSWEGym2
Dataset Description
This dataset is a filtered and deduplicated version of a merge containing many high quality SWE datasets, it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is specifically filtered for rows with complex/long code problems in the original datasets, having an average row size of 214.19kb, a total uncompressed size of 17.56GB, and a total of 85974 examples.
Dataset Details
Curated by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym2.DeepSWEGym2-Ultra
Dataset Description
This dataset is an EXTREMELY filtered and deduplicated version of DeepSWE-Gym2-Edu, it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is specifically filtered for rows with complex/long code problems in the original datasets, having an average row size of 8.27MB, a total uncompressed size of 8.27GB, and a total of 1000 examples.
Dataset Details
Curated by: MoreThought
Funded by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym2-Ultra.math-code-science-deepseek-r1-en
R1 Dataset Collection
Aggregated high-quality English prompts and model-generated responses from DeepSeek R1 and DeepSeek R1-0528.
Dataset Summary
The R1 Dataset Collection combines multiple public DeepSeek-generated instruction-response corpora into a single, cleaned, English-only JSONL file. Each example consists of a <|user|> prompt and a <|assistant|> response in one "text" field. This release includes:
~21,000 examples from the DeepSeek-R1-0528 Distilled Custom… See the full description on the dataset page: https://huggingface.co/datasets/Hugodonotexit/math-code-science-deepseek-r1-en.DeepSWEGym2-Edu
Dataset Description
This dataset is a filtered and deduplicated version of a merge containing many high quality SWE datasets, it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is specifically filtered for rows with complex/long code problems in the original datasets, having an average row size of 291.74kb, a total uncompressed size of 14.93GB, and a total of 53649 examples.
Dataset Details
Curated by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym2-Edu.DeepSWEGym2-Full
Dataset Description
This dataset is a merge containing many high quality SWE datasets, it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is NOT specifically filtered for rows with complex/long code problems in the original datasets, despite still having an average row size of 169.08kb, a total uncompressed size of 19.80GB, and a total of 122791 examples.
Dataset Details
Curated by: MoreThought
Funded by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym2-Full.Chinese-DeepSeek-R1-Distill-data-110k
中文基于满血DeepSeek-R1蒸馏数据集（Chinese-Data-Distill-From-R1）
🤗 Hugging Face | 🤖 ModelScope | 🚀 Github | 📑 Blog
注意：提供了直接SFT使用的版本，点击下载。将数据中的思考和答案整合成output字段，大部分SFT代码框架均可直接直接加载训练。
本数据集为中文开源蒸馏满血R1的数据集，数据集中不仅包含math数据，还包括大量的通用类型数据，总数量为110K。
为什么开源这个数据？
R1的效果十分强大，并且基于R1蒸馏数据SFT的小模型也展现出了强大的效果，但检索发现，大部分开源的R1蒸馏数据集均为英文数据集。 同时，R1的报告中展示，蒸馏模型中同时也使用了部分通用场景数据集。
为了帮助大家更好地复现R1蒸馏模型的效果，特此开源中文数据集。
该中文数据集中的数据分布如下：… See the full description on the dataset page: https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k.DeepSWEGym-Edu
Dataset Description
This dataset is a heavily filtered version of all the SWE-bench/SWE-smith-lang datasets (expect php) merged together (originally 88k rows total), it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is specifically filtered for rows with very complex/long code problems in the original datasets, having an average row size of 101.81kb, a total uncompressed size of 4.29GB, and 42529 examples total.… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym-Edu.DeepSWEGym-Full
Dataset Description
This dataset is a merged version of all the SWE-bench/SWE-smith-lang datasets (88k rows total), it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills.
It is NOT specifically filtered for rows with complex/long code problems in the original datasets, despite still having an average row size of 85.4kb, a total uncompressed size of 7.53GB, and 88130 examples total.
Dataset Details
Curated by:… See the full description on the dataset page: https://huggingface.co/datasets/MoreThought/DeepSWEGym-Full.Chinese-DeepSeek-R1-Distill-data-110k-SFT
中文基于满血DeepSeek-R1蒸馏数据集（Chinese-Data-Distill-From-R1）
🤗 Hugging Face | 🤖 ModelScope | 🚀 Github | 📑 Blog
注意：该版本为，可以直接SFT使用的版本，将原始数据中的思考和答案整合成output字段，大部分SFT代码框架均可直接直接加载训练。
本数据集为中文开源蒸馏满血R1的数据集，数据集中不仅包含math数据，还包括大量的通用类型数据，总数量为110K。
为什么开源这个数据？
R1的效果十分强大，并且基于R1蒸馏数据SFT的小模型也展现出了强大的效果，但检索发现，大部分开源的R1蒸馏数据集均为英文数据集。 同时，R1的报告中展示，蒸馏模型中同时也使用了部分通用场景数据集。
为了帮助大家更好地复现R1蒸馏模型的效果，特此开源中文数据集。该中文数据集中的数据分布如下：
Math：共计36568个样本，
Exam：共计2432个样本，
STEM：共计12648个样本，… See the full description on the dataset page: https://huggingface.co/datasets/Congliu/Chinese-DeepSeek-R1-Distill-data-110k-SFT.minimax-m3-deepsearchqa-skill-eval
MiniMax M3 DeepSearchQA Skill Eval
Evaluates minimax/minimax-m3 on google/deepsearchqa using a Pi agent, You.com MCP tools, and a research skill optimized for this harness, model, and tool surface.
MiniMax M3 Medium Reasoning with the You.com research skill reached 74.85% adjusted F1 on DeepSearchQA, above the paper's GPT-5 High Reasoning F1 result. Public artifacts are available for inspection and reproduction.
Links
GitHub:… See the full description on the dataset page: https://huggingface.co/datasets/youdotcom/minimax-m3-deepsearchqa-skill-eval.deepcontrol_output
DeepControl Output
This dataset contains trajectories running under the DeepControl inference pipeline across seven standard QA benchmarks.
DeepControl enables an LLM agent to actively regulate information acquisition during reasoning through (1) retrieval continuation control (when to keep searching), (2) granularity control (how much information to expand), and (3) annealed control training. Each record below is a full agent rollout: the model interleaves <think>, <search>… See the full description on the dataset page: https://huggingface.co/datasets/sxiong/deepcontrol_output.salabs-stem-deep-reasoning-cot-v13
🧪 SALabs Multi-Domain STEM Deep Reasoning & Chain-of-Thought (CoT) Corpus (v13.0)
[!IMPORTANT]
💳 Click Here to Purchase Enterprise Commercial License ($2,500 USD) & Instant 31.7MB Master Archive DownloadInstant download of the full lossless master package containing all 1,816 JSONL reasoning records + 13 complete uncompressed text corpora (31.72 MB uncompressed total) + commercial license certificate.
🌟 Executive Summary
The SALabs STEM Deep Reasoning… See the full description on the dataset page: https://huggingface.co/datasets/suitai/salabs-stem-deep-reasoning-cot-v13.DeepSeek-V4-Pro-Reasoning-8000x
DeepSeek-V4-Pro-Reasoning-8000x
This dataset contains 8,014 synthetic reasoning examples generated with DeepSeek V4 Pro through the DeepSeek API.
The release is branded as 8000x for readability, while the exact row count is 8,014.
This dataset is designed for supervised fine-tuning, reasoning distillation, and experimentation with long-form visible reasoning traces.
Dataset Summary
Release label: 8000x
Actual rows: 8,014
Teacher model: DeepSeek-V4-Pro… See the full description on the dataset page: https://huggingface.co/datasets/trjxter/DeepSeek-V4-Pro-Reasoning-8000x.DeepRethink
DeepRethink
Expanding AI thinking, more thinking needed
Thinking things and Contexts.hf-sanitized.hf-sanitized-BiKIfcWn9nbC8xiwxs370 .deeprethink-title { background-image: url('https://image.pollinations.ai/prompt/gradient%20dark%20and%20blue%20green%20bottom?width=1280&height=720&seed=2184&nologo=true&model=flux'); background-cover: bottom; -webkit-background-clip: text; background-clip: text; color: transparent; -webkit-text-fill-color: transparent; margin: 0 0 1rem 0; }… See the full description on the dataset page: https://huggingface.co/datasets/kulia-moon/DeepRethink.deepseek-v4-pro-math-cot-1k
DeepSeek V4 Pro Math CoT 1K
A small, high-signal supervised-fine-tuning (SFT) dataset of math reasoning traces. Problems were sampled from a Nemotron math problem set (originally sourced from StackExchange-Math and AoPS), answered by DeepSeek V4 Pro with thinking enabled at high reasoning effort, then independently reviewed by DeepSeek V4 Flash for correctness against the expected answer. Pathological reasoning traces (looping, run-away length, excessive Wait-style backtracking)… See the full description on the dataset page: https://huggingface.co/datasets/blythet/deepseek-v4-pro-math-cot-1k.ICBCBenchICBCBench: An Industry Consortium Benchmark for Financial Deep Research
Overview
ICBCBench is an industry consortium benchmark for evaluating financial Deep Research Agents in real-world research scenarios. It consists of bilingual objective and subjective tasks across major financial sectors, including capital markets, banking, insurance, and related financial services. Developed with over 50 contributors from more than 40 financial and academic organizations, ICBCBench… See the full description on the dataset page: https://huggingface.co/datasets/DeepFin-Intelligence/ICBCBench.CodeGen-Deep-5K
CodeGen-Deep-5K: Deep Reasoning for Competitive Programming
Part of the CodeGen suite | CodeGen-Diverse-5K (sister dataset)
Dataset Description
CodeGen-Deep-5K is a deep reasoning dataset designed for training code generation models with enhanced problem-solving capabilities. Unlike traditional datasets, this generates multiple distinct solutions for each problem, providing varied reasoning traces and approaches.
Key Statistics
Total samples: 5,000
Unique… See the full description on the dataset page: https://huggingface.co/datasets/Naholav/CodeGen-Deep-5K.hwtcm-deepseek-r1-distill-data
简介
DeepSeek蒸馏的传统中医数据集，原始数据来源于网络，未进行人工审查。
7B模型微调效果
模型表现出了推理能力，准确性有待继续验证。
我们的其他产品
中医NER：能识别方剂、本草、来源、病名、症状、证型，也许是基于BERT开源模型中识别最好的模型。中医考试题：也许是全网最早开源、数据最多的中医考试题，我们内部将其用于模型训练的性能评测数据集。中医SFT数据集：中医QA数据集，用于SFT微调。仓公：基于Qwen的指令微调模型（暂未开源）。仓公R1：基于DeepSeek蒸馏的超过100万条QA的指令微调模型，拥有强大的推理能力（暂未开源）。
。。。还有很多
Citation
If you find this project useful in your research, please consider cite:
@misc{hwtcm2024,
title={{hwtcm-deepseek-r1-distill-data} A traditional… See the full description on the dataset page: https://huggingface.co/datasets/Monor/hwtcm-deepseek-r1-distill-data.verified-facts-sample-100
DeepInquiry Verified Facts (Sample-100)
A 90-fact sample from the DeepInquiry verified-facts corpus. Every fact in this sample has been cross-checked against multiple structurally independent web sources, cited, dated, and confidence-scored before it entered the corpus.
This is a preview sample. The full corpus (~942 approved facts as of Sept 2026, growing continuously) is available via the DeepInquiry API at deepinquiry.ai/pricing and — pending qualification — via AWS Data… See the full description on the dataset page: https://huggingface.co/datasets/deepinquiry/verified-facts-sample-100.DeepSeek-R1-Distill-Qwen-32B-LeaPPaper: Learning from Peers in Reasoning Models
Project Page: https://learning-from-peers.github.io/
Code: https://github.com/tongxuluo/LeaP
sample-90
DeepInquiry Verified Facts (Sample-90)
A 90-fact sample from the DeepInquiry verified-facts corpus. Every fact in this sample has been cross-checked against multiple structurally independent web sources, cited, dated, and confidence-scored before it entered the corpus.
This is a preview sample. The full corpus (~942 approved facts as of Sept 2026, growing continuously) is available via the DeepInquiry API at deepinquiry.ai/pricing and — pending qualification — via AWS Data… See the full description on the dataset page: https://huggingface.co/datasets/deepinquiry/sample-90.Ko.SlimOrca원본 데이터셋: Open-Orca/SlimOrca
Chinese-DeepSeek-V3.2-Exp-chat-example
deepseek/deepseek-v3.2-exp (6.6K) 中文数据集样本
一、前言
本报告基于 deepseek/deepseek-v3.2-exp 模型（官方 API，8K 上下文窗口）进行数据集评测与可视化展示。测试数据集共包含 6,655 轮对话，语言覆盖以中文为主，辅以部分混合语种及非中文输入。本次报告旨在总结模型的对话特征、输入输出长度分布及上下文预算消耗情况，并为后续应用和优化提供参考。
二、数据与方法
数据来源：用户构建的 6,655 轮真实中文对话样本。
估算方法：
中文字符近似为 1 Token；
英文 4 字符 ≈ 1 Token；
用于规模与上下文预算对比，而非精确 Token 计数。
统计维度：
平均 Prompt/Output 长度（字符与估算 Token）；
总 Token 占上下文窗口比例；
语言分布（Prompt 语言类型）；
对话长度分布（用户提问、助手回答、总对话长度）。
三、总体结果
1. 样本概况… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/Chinese-DeepSeek-V3.2-Exp-chat-example.Ko.WizardLM_evol_instruct_V2_196k이 데이터셋은 자체 구축한 번역기로 WizardLM/WizardLM_evol_instruct_V2_196k을 번역한 데이터셋입니다. 아래 README 페이지도 번역기를 통해 번역되었습니다. 참고 부탁드립니다.
News
🔥 🔥 🔥 [08/11/2023] WizardMath 모델을 출시합니다.
🔥 WizardMath-70B-V1.0 모델은 ChatGPT 3.5, Claude Instant 1 및 PaLM 2 540B 를 포함 하 여 GSM8K에서 일부 폐쇄 소스 LLMs 보다 약간 더 우수 합니다.
🔥 우리의 WizardMath-70B-V1.0 모델은 SOTA 오픈 소스 LLM보다 24.8 포인트 높은 GSM8k Benchmarks에서 81.6 pass@1 을 달성합니다.
🔥 우리의 WizardMath-70B-V1.0 모델은 SOTA 오픈 소스 LLM보다 9.2 포인트 높은 MATH 벤치마크에서 22.7 pass@1 을 달성합니다.… See the full description on the dataset page: https://huggingface.co/datasets/nlp-with-deeplearning/Ko.WizardLM_evol_instruct_V2_196k.DeepShopdeepseek-v4-pro-pi-reasoning-sample-traces
DeepSeek V4 Pro Pi Reasoning Sample Traces
This dataset contains a compact sample of successful DeepSeek V4 Pro teacher trajectories for Pi-style reasoning and tool-use workflows.
It includes selected pass-only traces from these task providers:
abcbench
aider
autocodebench
bfcl
swebench
swtbench
termigen
Format
Each row contains:
id: stable sample id
segments: Qwen-style template-free supervised segments
label=false segments are context only
label=true segments… See the full description on the dataset page: https://huggingface.co/datasets/bytkim/deepseek-v4-pro-pi-reasoning-sample-traces.ko.SHP
🚢 Korean Stanford Human Preferences Dataset (Ko.SHP)
이 데이터셋은 자체 구축한 번역기를 활용하여 stanfordnlp/SHP 데이터셋을 번역한 것입니다.
아래의 내용은 해당 번역기로 README 파일을 번역한 것입니다. 참고 부탁드립니다.
If you mention this dataset in a paper, please cite the paper: Understanding Dataset Difficulty with V-Usable Information (ICML 2022).
Summary
SHP는 요리에서 법률 조언에 이르기까지 18가지 다른 주제 영역의 질문/지침에 대한 응답에 대한 385K 집단 인간 선호도 데이터 세트이다.
기본 설정은 다른 응답에 대 한 한 응답의 유용성을 반영 하기 위한 것이며 RLHF 보상 모델 및 NLG 평가 모델 (예: SteamSHP)을 훈련 하는 데… See the full description on the dataset page: https://huggingface.co/datasets/nlp-with-deeplearning/ko.SHP.ko.databricks-dolly-15k원본 데이터셋: databricks/databricks-dolly-15k
DeepTurtle
🐢 DeepTurtle: A Logic-Grounded Lateral Thinking Benchmark
DeepTurtle is a high-quality evaluation dataset designed to assess the Lateral Thinking and Implicit Logic Reasoning capabilities of Large Language Models (LLMs).
Unlike traditional QA datasets, DeepTurtle focuses on "Turtle Soup" puzzles (Situation Puzzles) where the model must deduce a complex hidden truth from a vague surface story. The dataset features a unique Logic Profile engine and contains real-world… See the full description on the dataset page: https://huggingface.co/datasets/YuiMax/DeepTurtle.DeepSeek-v3.1-reasoner-Distilled-math-samples
DeepSeek-V3.1 Distillation with NVIDIA Nemotron-Post-Training-Dataset-v2 (Math Subset)
The release of DeepSeek-V3.1 has attracted wide attention in the AI community. Its significant improvements in reasoning ability provide a new opportunity to explore optimization of domain-specific models. To investigate the potential of this model in complex mathematical reasoning tasks, I selected the math subset from NVIDIA’s newly released Nemotron-Post-Training-Dataset-v2 as seed problems and… See the full description on the dataset page: https://huggingface.co/datasets/Jackrong/DeepSeek-v3.1-reasoner-Distilled-math-samples.