Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,788 in the catalog so far, with new ones added every hour.

Data type

For which models

Find a dataset for your goal

Answered by Gemini 4 · the chat is kept only in this tab

Datasetmrlbenchmarks/global-piqa-nonparallelTop pick

Common sense benchmark for 136 language variants

Multiple-choice questions to test understanding of everyday logic and cultural features in rare languages.

  • Question answering
  • Multiple choice
  • acm
  • acq
  • Tunisian Arabic
  • +9
17.6K 4310K–100K rows
Datasetbowen-upenn/PersonaMem-v2

Dataset for LLM personalization and agent memory

10 to 100 thousand English examples for training models to work with hidden user profiles and memory.

  • Question answering
  • Text generation
  • English
17.3K 4010K–100K rows
Datasetopencsg/chinese-fineweb-edu-v2

Chinese Fineweb Edu V2 corpus

Large Chinese pre-training corpus; version V2 is marked as deprecated.

  • Text generation
  • Chinese
16K 75100M–1B rows
Datasetlmsys/mt_bench_human_judgments

MT-bench Human Judgment Dataset

3.3k expert pairwise preference annotations for six model responses to 80 MT-bench questions.

  • Question answering
  • English
15.8K 1501K–10K rows
DatasetPKU-Alignment/PKU-SafeRLHF

Preferences and Q-A for LLM safety

Data for training and evaluating safe responses: preferences, question-answer pairs, and English prompts.

  • Text generation
  • English
14.9K 207100K–1M rows
Datasetbigcode/the-stack-dedup

Deduplicated source code corpus

Code in many programming languages with permissive licenses after near-deduplication, about 1.5 TB.

  • Text generation
  • Code
14.1K 431100M–1B rows🔒 on request
Datasetkotekjedi/stolen-thoughtsTop pick

Recovered hidden reasoning of 27 closed LLMs

Dataset for research on extracting hidden reasoning chains from proprietary APIs on complex tasks.

  • Text generation
  • English
13.6K 810K–100K rows🔒 on request
DatasetHAERAE-HUB/KMMLU

Korean KMMLU benchmark

35,030 expert multiple-choice questions across 45 subjects in Korean.

  • Multiple choice
  • Korean
13.4K 106100K–1M rows
Datasetstanfordnlp/SHP

385K human response preferences

385K human preferences for answer helpfulness across 18 topics.

  • Text generation
  • Question answering
  • English
13.1K 333100K–1M rows
Datasetmikex86/stackoverflow-posts

Archive of 60 million StackOverflow posts in Markdown

A large corpus of StackOverflow questions, answers, and code up to June 14, 2023, in Markdown text.

  • Question answering
  • Text generation
  • Code
  • English
13.1K 6410M–100M rows
DatasetCohereLabs/aya_collection

Multilingual Aya collection for NLP tasks

A large multilingual collection of texts for classification, summarization, and translation.

  • Text classification
  • Summarization
  • Translation
  • Acehnese
  • Afrikaans
  • Amharic
  • +9
11.9K 252100M–1B rows
DatasetBAAI/Infinity-Instruct

Infinity-Instruct: a large instruction dataset

Millions of English and Chinese instructions for LLM fine-tuning and evaluation.

  • Text generation
  • English
  • Chinese
11.7K 77110M–100M rows🔒 on request
DatasetSimpleStories/SimpleStories

Over 2 million short English stories

A synthetic corpus of simple stories for training small and interpretable language models.

  • Text generation
  • English
11K 441M–10M rows
Datasetargilla/ultrafeedback-binarized-preferences-cleaned

UltraFeedback binarized preferences

Cleaned preference pairs from UltraFeedback for language model fine-tuning.

  • Text generation
  • English
10.7K 17610K–100K rows
DatasetTeichAI/DeepSeek-v4-Pro-Agent

DeepSeek v4 Pro Agent Traces

4,006 raw agent traces with DeepSeek v4 Pro responses and a tool schema snapshot for SFT.

  • Text generation
  • English
9.6K 1111K–10K rows
Datasetairtrain-ai/fineweb-edu-fortified

Cleaned FineWeb-Edu with embeddings

Large English FineWeb-Edu corpus after precise deduplication, featuring embeddings and row repetition counters.

  • Text generation
  • English
9.2K 71100M–1B rows
Datasetclue/clue

CLUE Chinese language understanding benchmark

Text and tabular tasks for training and evaluating Chinese language understanding models.

  • Text classification
  • Multiple choice
  • Chinese
7.6K 46100K–1M rows
Datasetargilla/distilabel-capybara-dpo-7k-binarized

7k-example Capybara DPO dataset

Binarized preferences for DPO based on Capybara: multi-turn dialogues for training chat models.

  • Question answering
  • Text generation
  • English
7.5K 1931K–10K rows
Datasetopenbmb/UltraData-Stack-v3-L2Top pick

Cleaned 240B-token source code corpus

Filtered code corpus of ~240B tokens across 11 programming languages from The Stack v3.

  • Text generation
  • English
  • Chinese
7.4K 37100M–1B rows
Datasetnvidia/Nemotron-ClimbMix

High-quality corpus for LLM pretraining

An English dataset for LLM pretraining focused on high quality and efficient scaling.

  • Text generation
  • English
7.1K 152100M–1B rows
Datasetdell-research-harvard/newswire

Historical US wire service news

2.7 million US wire service articles from 1878-1977 with geolocation, topics, and entities.

  • Text classification
  • Text generation
  • Text retrieval
  • English
6.9K 951M–10M rows