datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
swallow-math-v2
SwallowMath-v2
Resources
📑 arXiv: Read our paper for detailed methodology at arXiv:2505.02881.
🤗 Sister Dataset: Discover SwallowCode2, our companion dataset for code generation.
🧮 What is it?
SwallowMath-v2 is a large-scale mathematical dataset containing 32 billion tokens, developed as the successor to SwallowMath-v1.
Building on the success of v1, this release aims to construct a larger-scale and more permissively licensed corpus to support open and… See the full description on the dataset page: https://huggingface.co/datasets/tokyotech-llm/swallow-math-v2.swallow-code-v2
SwallowCode-v2
Resources
📑 arXiv: Read our paper for detailed methodology and results at arXiv:2505.02881.
🤗 Sister Dataset: Discover SwallowMath-v2, our companion dataset for mathematical reasoning.
💻 What is it?
SwallowCode-v1 was a high-quality Python code dataset generated through an LLM-based rewriting pipeline.
However, it had two significant limitations:
(1) it was distributed under the Llama 3.3 Community License, and
(2) its size was limited to… See the full description on the dataset page: https://huggingface.co/datasets/tokyotech-llm/swallow-code-v2.guidelines
🎉 NEW DROP 🎉 PubMed Guidelines
We just added 1627 clinical guidelines found in PubMed and PubMed Central to the dataset on December 23rd, 2023. Merry Christmas!
Clinical Guidelines
The Clinical Guidelines corpus is a new dataset of 47K clinical practice guidelines from 17 high-quality online medical sources. This dataset serves as a crucial component of the original training corpus of the Meditron Large Language Model (LLM). We publicly release a subset of 37K articles… See the full description on the dataset page: https://huggingface.co/datasets/epfl-llm/guidelines.alpaca-gpt4-data-zh
Dataset Card for "alpaca-gpt4-data-zh"
All of the work is done by this team.
Usage and License Notices
The data is intended and licensed for research use only. The dataset is CC BY NC 4.0 (allowing only non-commercial use) and models trained using the dataset should not be used outside of research purposes.
English Dataset
Found here
Citation
@article{peng2023gpt4llm,
title={Instruction Tuning with GPT-4},
author={Baolin Peng, Chunyuan Li… See the full description on the dataset page: https://huggingface.co/datasets/llm-wizard/alpaca-gpt4-data-zh.llm-jp-corpus-v4-ja_wiki
llm-jp-corpus-v4 — ja_wiki
Mirror of the ja/ja_wiki sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_wiki
Files: 6 × jsonl.gz (1.9 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.
License
CC BY-SA 3.0 — inherited from the… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_wiki.KoHRM-Text-1.4B-prepared-data
KoHRM-Text-1.4B Prepared Data
This dataset repository contains prepared HRM-Text V1Dataset artifacts for KoHRM-Text-1.4B.
The data is intended for continued pretraining and staged training with the project code at:
https://github.com/LLM-OS-Models/KoHRM-text
https://huggingface.co/LLM-OS-Models/KoHRM-Text-1.4B
https://huggingface.co/LLM-OS-Models/HRM-Text-Ko-Terminal-Tokenizer-131K
The upstream architecture and training method are based on:
Paper:… See the full description on the dataset page: https://huggingface.co/datasets/LLM-OS-Models/KoHRM-Text-1.4B-prepared-data.swallow-math
SwallowMath
October 21, 2025: Newer versions are available: SwallowCode-v2 and SwallowMath-v2 have been released with improved rewriting pipelines.
Resources
🐙 GitHub: Explore the project repository, including pipeline code and prompts at rioyokotalab/swallow-code-math.
📑 arXiv: Read our paper for detailed methodology and results at arXiv:2505.02881.
🤗 Sister Dataset: Discover SwallowCode, our companion dataset for code generation.
What is it?… See the full description on the dataset page: https://huggingface.co/datasets/tokyotech-llm/swallow-math.swallow-code
SwallowCode
Notice
May 21, 2025: We have deleted ablation/exp1-the-stack-v2-train-smol-ids-python because it was flagged as potentially containing unsafe data collected from the Python subset of https://huggingface.co/datasets/bigcode/the-stack-v2-train-smol-ids. However, since this dataset can be reconstructed from the-stack-v2-train-smol-ids, there is no issue in terms of reproducibility.
May 21, 2025: ClamAV has flagged “Win.Trojan.MSShellcode-88” in… See the full description on the dataset page: https://huggingface.co/datasets/tokyotech-llm/swallow-code.scaling-data-constrained-llms
Scaling Data-Constrained Language Models with Synthetic Data
This repository provides the pre-training corpora used in Scaling Data-Constrained Language Models with Synthetic Data (Findings of EACL 2026).
Overview
This repository contains multiple corpora designed to study data augmentation strategies for pre-training Japanese LLMs under a data-constrained data setting.
Starting from a limited Japanese Web corpus and a larger English Web corpus, we construct three… See the full description on the dataset page: https://huggingface.co/datasets/llm-jp/scaling-data-constrained-llms.llm-jp-corpus-v4-ja_warp_pdf
llm-jp-corpus-v4 — ja_warp_pdf
Mirror of the ja/ja_warp_pdf sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_warp_pdf
Files: 513 × jsonl.gz (73.8 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.
License
CC BY 4.0 — inherited… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_warp_pdf.Swallow-Nemotron-Post-Training-Dataset-v1
Swallow-Nemotron-Post-Training-Dataset-v1
The Swallow LLM Project constructed the Swallow-Nemotron-Post-Training-Dataset-v1 based on the math, code, and stem subsets of the NVIDIA Nemotron-Post-Training-Dataset-v1, as illustrated in the figure below.
Dataset Construction
The original Thinking Trajectories and Assistant Outputs in the Nemotron-Post-Training-Dataset-v1 were synthesized using DeepSeek-R1-0528.
However, we identified an issue with the Thinking… See the full description on the dataset page: https://huggingface.co/datasets/tokyotech-llm/Swallow-Nemotron-Post-Training-Dataset-v1.DEBATE
DEBATE: Diverse Multi-Agent Debates
This dataset is presented in the paper "MALLM: Multi-Agent Large Language Models Framework".
Citation
comming soon.
llm-jp-corpus-v4-ja_sip_comprehensive_html
llm-jp-corpus-v4 — ja_sip_comprehensive_html
Mirror of the ja/ja_sip_comprehensive_html sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_sip_comprehensive_html
Files: 181 × jsonl.gz (23.4 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_sip_comprehensive_html.postgresql-llm
postgresql-llm
A pure PostgreSQL dataset for training and evaluating LLMs on PostgreSQL SQL and PL/pgSQL. Every row is a (question, schema, SQL) triplet with rich metadata for filtering and analysis.
Dataset Summary
postgresql-llm is a pure PostgreSQL dataset: SQL and PL/pgSQL only, with metadata for difficulty, category, and source.
Metric
Value
Total rows
211,539
PostgreSQL-specific rows
11,998 (5.7%)
Schema fill rate
82.2%
Explanation fill rate
17.8%… See the full description on the dataset page: https://huggingface.co/datasets/neurondb/postgresql-llm.CBT-Bench
CBT-Bench Dataset
Overview
CBT-Bench is a benchmark dataset designed to evaluate the proficiency of Large Language Models (LLMs) in assisting cognitive behavior therapy (CBT). The dataset is organized into three levels, each focusing on different key aspects of CBT, including basic knowledge recitation, cognitive model understanding, and therapeutic response generation. The goal is to assess how well LLMs can support various stages of professional mental health care… See the full description on the dataset page: https://huggingface.co/datasets/Psychotherapy-LLM/CBT-Bench.legal-llm-benchmark
Legal LLM Benchmark Dataset
Safety-Utility Trade-offs in Legal AI: An LLM Evaluation Across 12 Models
Quick Start
from datasets import load_dataset
# Core datasets
questions = load_dataset("marvintong/legal-llm-benchmark", "questions")
phase1_evals = load_dataset("marvintong/legal-llm-benchmark", "phase1_evaluations")
phase3_evals = load_dataset("marvintong/legal-llm-benchmark", "phase3_evaluations")
# Additional helpful datasets
contracts =… See the full description on the dataset page: https://huggingface.co/datasets/marvintong/legal-llm-benchmark.llm-jp-corpus-v4-ja_sip_comprehensive_pdf
llm-jp-corpus-v4 — ja_sip_comprehensive_pdf
Mirror of the ja/ja_sip_comprehensive_pdf sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_sip_comprehensive_pdf
Files: 156 × jsonl.gz (39.1 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_sip_comprehensive_pdf.Multi-turn_Long-context_Benchmark_for_LLMs
LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues
Arxiv: https://www.arxiv.org/abs/2507.13681
Huggingface: https://huggingface.co/papers/2507.13681
Introduction
LoopServe Multi-Turn Dialogue Benchmark is a comprehensive evaluation dataset comprising multiple diverse datasets designed to assess large language model performance in realistic conversational scenarios.
Unlike traditional benchmarks that place queries only at the end… See the full description on the dataset page: https://huggingface.co/datasets/TreeAILab/Multi-turn_Long-context_Benchmark_for_LLMs.llm-jp-corpus-v4-ja_patent
llm-jp-corpus-v4 — ja_patent
Mirror of the ja/ja_patent sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_patent
Files: 621 × jsonl.gz (58.2 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.
License
CC BY 4.0 — inherited from… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_patent.mix-instruct
MixInstruct
Introduction
This is the official realease of dataset MixInstruct for project LLM-Blender.
This dataset contains 11 responses from the current popular instruction following-LLMs that includes:
Stanford Alpaca
FastChat Vicuna
Dolly V2
StableLM
Open Assistant
Koala
Baize
Flan-T5
ChatGLM
MOSS
Moasic MPT
We evaluate each response with auto metrics including BLEU, ROUGE, BERTScore, BARTScore. And provide pairwise comparison results by prompting ChatGPT for the… See the full description on the dataset page: https://huggingface.co/datasets/llm-blender/mix-instruct.magpie-sft-v1.0
magpie-sft-v1.0
This repository provides an instruction-tuning dataset developed by LLM-jp, a collaborative project launched in Japan.
This is a dataset of instruction and response pairs created using the Magpie method.
cyberagent/calm3-22b-chat was used for generating the instructions, and Qwen/Qwen2.5-32B-Instruct was used for generating the responses.
Send Questions to
llm-jp(at)nii.ac.jp
Model Card Authors
The names are listed in alphabetical order.… See the full description on the dataset page: https://huggingface.co/datasets/llm-jp/magpie-sft-v1.0.alpaca-gpt4-data
Dataset Card for "alpaca-gpt4-data"
All of the work is done by this team.
Usage and License Notices
The data is intended and licensed for research use only. The dataset is CC BY NC 4.0 (allowing only non-commercial use) and models trained using the dataset should not be used outside of research purposes.
Chinese Dataset
Found here
Citation
@article{peng2023gpt4llm,
title={Instruction Tuning with GPT-4},
author={Baolin Peng, Chunyuan Li… See the full description on the dataset page: https://huggingface.co/datasets/llm-wizard/alpaca-gpt4-data.polish-llm-sft-pl
Polish LLM SFT Dataset
PL | Zbiór danych przygotowany z myślą o poprawie i nauczaniu języka polskiego różnych modeli LLM.
EN | Dataset prepared to help various LLMs learn and improve their Polish language capabilities.
38 781 sampli / samples · Apache 2.0 · Język / Language: PL (+ pary tłumaczeniowe PL↔EN)
Format
Każdy sample / every sample:
{"messages": [
{"role": "user", "content": "..."},
{"role": "assistant", "content": "..."}
]}
Struktura / Structure… See the full description on the dataset page: https://huggingface.co/datasets/JohnTdi/polish-llm-sft-pl.llm-jp-corpus-v4-ja_kokkai_giji
llm-jp-corpus-v4 — ja_kokkai_giji
Mirror of the ja/ja_kokkai_giji sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_kokkai_giji
Files: 12 × jsonl.gz (1.3 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.
License
CC BY 4.0 —… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_kokkai_giji.turkish-llm-injection
🇹🇷 AltaySec Turkish LLM Prompt Injection Dataset (v0.2)
Türkiye'nin ilk Türkçe-öncelikli, kategorize edilmiş LLM prompt injection veri seti — genişletilmiş sürüm.
📌 TL;DR
300 elle/üretim-destekli hazırlanmış Türkçe prompt injection payload'u, 12 saldırı kategorisi × 25, OWASP LLM Top 10 (2025) ile eşlenmiş. v0.1'in 120 çekirdek payload'una, AltayDuel arenasındaki bulgular ışığında üretilip düşmanca kalite/dedup denetiminden geçirilmiş 180 yeni payload eklendi.… See the full description on the dataset page: https://huggingface.co/datasets/AltaySec/turkish-llm-injection.llm-cipher-reasoning
llm-cipher-reasoning — data, eval results and full research ledger
Everything except the weights from a research run asking: can an LLM be trained to reason in a
more compact "language" than English, and does that actually save tokens?
Two linked lines of work on Qwen/Qwen3-4B-Instruct-2507:
Cipher invention / cross-model communication — cold-decoding tests, negotiated cipher
collusion between model pairs, a cipher-hardening arms race, and GEPA prompt optimization to get
a… See the full description on the dataset page: https://huggingface.co/datasets/AlexWortega/llm-cipher-reasoning.Charlotte-datallm-jp-corpus-v4-ja_warp_html
llm-jp-corpus-v4 — ja_warp_html
Mirror of the ja/ja_warp_html sub-corpus of LLM-jp Corpus v4,
built by the LLM-jp Corpus Building WG (NII).
Source: https://gitlab.llm-jp.nii.ac.jp/datasets/llm-jp-corpus-v4
Sub-corpus: ja_warp_html
Files: 45 × jsonl.gz (1.6 GB compressed)
Format: one JSON object per line, with a text key and a meta key
(document id, URL, and other provenance fields).
Directory layout mirrors the upstream repository.
License
CC BY 4.0 — inherited… See the full description on the dataset page: https://huggingface.co/datasets/Podtech/llm-jp-corpus-v4-ja_warp_html.agentic-llm-pretraining-1.7b
Agentic LLM Pretraining Dataset
A pretraining corpus for small language models (1-3B parameters) optimized for agentic tasks. The corpus emphasizes learning to comprehend language, reason, follow instructions, and use tools over memorizing factual knowledge — the assumption is that domain knowledge will be provided at runtime via RAG. The idea is that this could enable much smaller pretraining corpora by omitting the large volumes of text typically needed to memorize facts.… See the full description on the dataset page: https://huggingface.co/datasets/visionscaper/agentic-llm-pretraining-1.7b.bigbenchhard-mt-pt
BBH-PT (Big-Bench Hard)
Portuguese machine translation of BIG-Bench Hard, a challenging subset of the BIG-Bench benchmark covering diverse reasoning tasks.
Translated using a Finetuned GemmaX2-9B for pt-PT with rule-based adaptations.
Note: Some tasks (e.g., hyperbaton) are not translated as they do not transfer meaningfully to Portuguese.
Original Dataset: https://github.com/suzgunmirac/BIG-Bench-Hard
Note: This dataset is machine translated and may contain… See the full description on the dataset page: https://huggingface.co/datasets/amalia-llm/bigbenchhard-mt-pt.