Skip to content

Datasets

Hugging Face datasets with plain-language descriptions: what's inside and what models to train on them. 4,797 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

DatasetVoidreaper2026/cybersec-master-dataset

Large cybersecurity instructional corpus

Over 1.8 million cleaned English dialogues in ShareGPT format for training LLMs on information security tasks.

  • Text generation
  • Question answering
  • English
509 651M–10M rows🔒 on request
DatasetFINAL-Bench/MetacognitiveTop pick

LLM metacognitive benchmark

Testing whether LLMs know what they do not know and can correct their mistakes.

  • Text generation
  • Question answering
  • English
507 110up to 1K rows
Datasetbeyoru/Aesir-Character-CoT-roleplay

Character-persona reasoning roleplay dialogues

1,973 English roleplay dialogues with chains of thought (CoT) written in character.

  • Text generation
  • English
499 531K–10K rows
Datasetbigcode/the-stack-github-issues

GitHub Issues and Pull Requests discussions (54 GB)

A corpus of 30.9 million files featuring developer discussions in GitHub issues and pull requests.

  • Text generation
  • English
492 5310M–100M rows🔒 on request
DatasetFutureMa/DramaBench

DramaBench Script Continuation Benchmark

1,103 context-continuation pairs to evaluate LLMs' ability to continue dramatic scenarios.

  • Text generation
  • English
488 1111K–10K rows
Datasetcorbyrosset/researchy_questions

Complex user research queries from Bing

Around 100 thousand complex multi-aspect English questions requiring deep search and analysis.

  • Question answering
  • English
487 3810K–100K rows
Datasetalespalla/chatbot_instruction_prompts

Curated corpus of 500 thousand instructions and answers

A combined and artifact-cleaned set of English prompts and answers for training instruct models.

  • Question answering
  • Text generation
  • English
486 66100K–1M rows
Datasetankitjh4/bharat-government-documents

Indian government reference documents

A cleaned corpus of official public Indian documents in parquet format for search and question-answering systems.

  • Question answering
  • Text retrieval
476 6510K–100K rows
Datasetxingyaoww/code-act

Executable code agent instructions

CodeActInstruct data: multi-step tasks where the agent acts via executable Python code.

  • Text generation
  • English
472 7710K–100K rows
Datasetacon96/Home-Assistant-Requests

Smart home control commands and responses

English prompt-response pairs for training an assistant to work with Home Assistant.

  • Question answering
  • Text generation
  • English
472 4810K–100K rows
Datasetmii-llm/gazzetta-ufficiale

Official Italian laws and legal acts

Millions of texts from the Italian official gazette Gazzetta Ufficiale for training language models.

  • Text generation
  • Fill mask
  • Italian
472 421M–10M rows
DatasetIlyaGusev/gpt_roleplay_realm

GPT roleplay realm

GPT-4 characters and dialogues in Russian and English for training roleplay language models.

  • Text generation
  • Russian
  • English
469 105up to 1K rows
Datasetsoberania/jabuticabaTop pick

Largest cleaned Portuguese corpus for LLMs

669 GB of cleaned and deduplicated Portuguese texts (over 139 billion tokens) for LLM training.

  • Text generation
  • Fill mask
  • Portuguese
467 25100M–1B rows🔒 on request
DatasetFlmc/DISC-Med-SFT

Chinese medical QA dataset from DISC-Med-SFT

A subset of DISC-Med-SFT with Chinese text medical question-answer pairs for training and evaluating QA models.

  • Question answering
  • Chinese
466 101100K–1M rows
DatasetAlicanKiraz0/Turkish-SFT-Dataset-v1.0

Turkish SFT dataset with long responses

Around 5,500 Turkish dialogue examples with long outputs for SFT.

  • Text classification
  • Question answering
  • Text generation
  • Turkish
466 531K–10K rows
DatasetAI-companionship/INTIMATop pick

AI dialogue attachment and empathy benchmark

English evaluation set for testing how language models respond to users' emotional requests.

  • Text generation
  • English
463 38up to 1K rows
DatasetGaryYang123/zh-meme-sft-8k

Chinese internet meme SFT dataset

Around 8k real meme dialogues from Douyin, Xiaohongshu, and Bilibili for conversational model training.

  • Text generation
  • Chinese
457 881K–10K rows
DatasetPageshift-Entertainment/LongPageTop pick

Dataset for generating novels with outlines

Long novels and hierarchical outlines for training models in coherent storytelling.

  • Text generation
  • English
455 1661K–10K rows
Datasetturkish-nlp-suite/InstrucTurca

Synthetic Instruction Dataset in Turkish

1 million to 10 million examples for instruction fine-tuning of Turkish language models across various topics.

  • Text generation
  • Question answering
  • Turkish
446 421M–10M rows