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

Datasetbench-llm/or-bench

OR-Bench language model over-refusal benchmark

A prompt set to check the balance between LLM safety and their tendency to refuse harmless prompts.

  • Text generation
  • Question answering
  • English
21.7K 2610K–100K rows
Datasetbastao/VeraCruz_PT-BR

Large Portuguese Text Corpus

Approximately 190 million Portuguese texts divided into European and Brazilian variants.

  • Text generation
  • Text classification
  • Portuguese
21.6K 18100M–1B rows
Datasetmlfoundations/MINT-1T-PDF-CC-2024-10Top pick

Multimodal PDF corpus for VLM pre-training

Part of the giant open MINT-1T dataset featuring interleaved English text and PDF images.

  • Image → text
  • Text generation
  • English
21.5K 51M–10M rows
Datasetyongchao98/SymBench

Symbolic tasks and data for SFT and DPO

A dataset of 37 symbolic tasks, generation scripts, and English data for SFT and DPO training.

  • Text generation
  • English
21.4K 51K–10K rows
Datasetnvidia/Nemotron-Pretraining-Indic-Specialized-v1Top pick

Nemotron-Pretraining-Indic Corpus for Indian Languages

A massive specialized text corpus from NVIDIA containing 100M to 1B records in Indian languages for language model pre-training.

  • Text generation
  • Assamese
  • Bangla
  • Bodo
  • +9
21.3K 7100M–1B rows
Datasetfka/prompts.chat

Community prompts for AI chats

Mirror of an open platform with user prompts for ChatGPT and other AI chats.

  • Question answering
  • Text generation
21.3K 10K1K–10K rows
Datasetfuturehouse/lab-bench

Biological research benchmark

A set of text and visual tasks to evaluate AI agent capabilities in scientific biology.

  • Question answering
21.3K 541K–10K rows
DatasetNousResearch/hermes-function-calling-v1

Function calling and JSON-mode dataset

Dialogues for training LLMs in function calling, JSON mode, and structured extraction.

  • Text generation
  • Question answering
  • Embeddings
  • English
20.9K 51310K–100K rows
DatasetHuggingFaceFW/finepdfs-eduTop pick

Educational PDF Tokens in 69 Languages

350B+ educational PDF tokens filtered by an educational quality classifier.

  • Text generation
  • English
  • German
  • Japanese
  • +9
20.9K 10710M–100M rows
DatasetFreedomIntelligence/medical-o1-reasoning-SFT

Medical reasoning data for SFT

Medical tasks with chain-of-thought reasoning in English and Chinese for SFT and reasoning model initialization.

  • Question answering
  • Text generation
  • English
  • Chinese
20.8K 1.2K10K–100K rows
Datasetlikaixin/InstructCoder

114k examples for training AI code editing

English dataset of instruction-source-edit triplets for adapting language models to code modification.

  • Text generation
  • English
20.6K 18100K–1M rows
Datasetmlfoundations/MINT-1T-PDF-CC-2023-23Top pick

MINT-1T multimodal corpus chunk (PDF and web data)

Part of a giant open dataset with interleaved texts and images from PDFs and the web.

  • Image → text
  • Text generation
  • English
20.6K 101M–10M rows
Datasetakshathmangudi/SciCode

SciCode scientific code generation benchmark

Complex English text tasks for evaluating the ability of language models to write code for scientific research.

  • Question answering
  • English
20.4K 01K–10K rows
Datasetopenbmb/Ultra-FineWeb-L3

Ultra-FineWeb-L3 cleaned web corpus

High-quality English and Chinese L3 web texts with enhanced reasoning signals for LLM training.

  • Text generation
  • English
  • Chinese
20.2K 3471B–10B rows
Datasetinstruction-pretrain/general-instruction-augmented-corporaTop pick

200M instruction-response pairs for LLM pre-training

Massive English corpus of 200M instruction-response pairs across 40+ task categories for multi-task model pre-training.

  • Text classification
  • Table question answering
  • Question answering
  • English
20.1K 31
DatasetHumanBehaviorAtlas/human_behavior_atlasTop pick

Multimodal dataset for behavior and emotion analysis

Combined corpus of 16 datasets with video, audio, and text for recognizing emotions, humor, and human state.

  • Video classification
  • Audio classification
  • Text classification
  • English
  • Chinese
20K 4100K–1M rows
DatasetSakanaAI/ALE-BenchTop pick

Competitive programming benchmark from Sakana AI and AtCoder

Official set of tasks from AtCoder contests for evaluating AI systems in algorithmic programming with scoring.

  • Image text → text
  • Reinforcement learning
  • Text generation
  • English
  • Japanese
20K 13up to 1K rows
Datasetheitorrosa/cvm-corpus

Brazilian regulator CVM financial report corpus

Cleaned and chunked documents from Brazilian regulator CVM in Portuguese for language model fine-tuning.

  • Text generation
  • Fill mask
  • Portuguese
20K 0100K–1M rows
DatasetChrisneverdie/OnlySports_Dataset

Large-scale English sports text corpus

1 to 10 billion English sports documents for training specialized language models.

  • Text generation
  • English
20K 61B–10B rows
DatasetZomiLearner/English-Zomi-OPUS_Tatoeba_v20230412Top pick

1.78M phrase parallel corpus: English and Zomi

An open bilingual dataset of 1.78 million sentence pairs for the rare Zomi (Tedim Chin) language and English.

  • Translation
  • English
  • ctd
19.9K 11M–10M rows
DatasetAzzindani/Legal_Corpus_QA_SynDeepThink

Indonesian legal question and answer dataset

A synthetic dataset with Q&A pairs and step-by-step logic verification for legal language models.

  • Text generation
  • Question answering
  • Indonesian
19.9K 11K–10K rows