Synthetic reasoning and web search dataset for small models
A 700+ million token corpus for training compact models in step-by-step task analysis and web search.
- Text generation
- Question answering
- Multimodal
- English
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.
Answered by Gemini 4 · the chat is kept only in this tab
A 700+ million token corpus for training compact models in step-by-step task analysis and web search.
The first benchmark for evaluating recognition and parsing of real digital and photographed documents across 6 languages.
An open scenario from Artificial Analysis for testing AI agent performance on complex work tasks.
Tables with a list of scientific publications from The Prompt Report review for reproducing the systematic analysis.
160k synthetic and 200k augmented examples of user, assistant, and tool interactions.
Test tasks for evaluating autonomous agents in healthcare scenarios: from treatment authorization to patient management.
From 1 to 10k document pages for comparing OCR and vision-language models across three tracks.
A curated selection of English books on AI, ML, DL, RL, transformers, and LLMs for structured study.
Large web dataset for detecting and classifying input fields on document scans and pages.
A small collection of documents with material removal requests received by the Hugging Face team.
49,662 real mechanical engineering samples in four formats: PDF drawing, PNG image, STEP 3D model, and CadQuery code.
Massive corpus of Turkish parliamentary transcripts in PDF format for long-text processing.
Benchmark for evaluating the accuracy of structured data extraction from resumes by JSON Schema.
Synthetic business PDFs and nearly 10,000 questions on tables, charts, forms, and text for evaluating Document QA systems.
English and Chinese PDF documents for testing scientific literature metadata extraction.
75 multi-page FCC invoices with JSON schemas for structured data extraction.
A collection of research papers, code, chats, and traces from autonomous scientific agents on a preprint server.
Complete public archive of Reserve Bank of India (RBI) regulatory PDF documents from 1991 to 2026.
Around 1,500 Chinese-translated light novels in text format and PDFs with illustrations for creative text generation.
An initiative to unify over 1,000 open medical imaging and document datasets.
Testing whether LLMs know what they do not know and can correct their mistakes.
English evaluation set for testing how language models respond to users' emotional requests.
Annotated restaurant reviews in Uzbek for aspect extraction and sentiment determination.
The first benchmark based on 17,600 real veterinary records annotated with ICD-11 disease codes.