--- language: - afr - amh - ara - arz - aze - bel - ben - bul - cat - ces - cym - dan - deu - ell - eng - est - eus - fas - fin - fra - gle - glg - guj - hbs - heb - hin - hrv - hun - hye - ind - isl - ita - jpn - kan - kat - kaz - khm - kir - kor - lav - lit - ltz - mal - mar - mkd - mlt - mon - msa - mya - nep - nld - nor - pan - pol - por - ron - rus - sin - slk - slv - spa - sqi - srp - swe - tam - tel - tgl - tha - tur - ukr - urd - uzb - vie - yid - zho multilinguality: - multilingual task_categories: - question-answering config_names: - afr - amh - ara - arz - aze - bel - ben - bul - cat - ces - cym - dan - deu - ell - eng - est - eus - fas - fin - fra - gle - glg - guj - hbs - heb - hin - hrv - hun - hye - ind - isl - ita - jpn - kan - kat - kaz - khm - kir - kor - lav - lit - ltz - mal - mar - mkd - mlt - mon - msa - mya - nep - nld - nor - pan - pol - por - ron - rus - sin - slk - slv - spa - sqi - srp - swe - tam - tel - tgl - tha - tur - ukr - urd - uzb - vie - yid - zho tags: - question-answering license: cc-by-4.0 size_categories: - 10MWebFAQ Q&A Dataset

Overview | Details | Structure | Examples | Considerations | License | Citation | Contact | Acknowledgement

## Overview The **WebFAQ Q&A Dataset** is a broad-coverage corpus of **96 million** natural question-answer (QA) pairs in **75 languages**, gathered from FAQ pages on the web. It leverages structured [schema.org FAQPage](https://schema.org/FAQPage) annotations, making it a unique resource for large-scale Question Answering research. Each entry includes a question, the corresponding answer, and additional metadata such as language, topic, and question type. **Why is it useful?** - Designed for **multilingual open-domain Question Answering** research. - Offers substantial linguistic diversity (49% non-English samples). - Provides labeled metadata for deeper analyses (topic, question type). **Background** - Data was collected from the [Common Crawl](https://commoncrawl.org) dumps (Oct 2022 – Oct 2024) processed by [Web Data Commons](https://webdatacommons.org). - After parsing and cleaning, the final corpus contains QAs that are intended by website owners to be publicly accessible. The WebFAQ collection further includes: 1. **Retrieval Datasets** If you need a curated subset of WebFAQ for **training or benchmarking retrieval models**, check out the [WebFAQ Retrieval Datasets](https://huggingface.co/datasets/PaDaS-Lab/webfaq-retrieval), which include official train/test splits in 20 languages. 2. **Bilingual Corpora** WebFAQ also provides **QA-aligned bilingual datasets** (bitexts) created via state-of-the-art cross-lingual sentence embedding and automated translation evaluation. These can be used for **machine translation, bitext mining**, or **cross-lingual IR** experiments. Please check out the check out the [WebFAQ Bilingual Datasets](https://huggingface.co/datasets/PaDaS-Lab/webfaq-bitexts) for more information. ## Details ### Languages 75 languages with ≥1,000 samples each, including low-resource languages. 49 of these languages appear in ≥100 distinct websites, ensuring variety in topics and writing style. | Language | % | Language | % | |----------------|------|----------------|------| | eng (English) | 51.2 | ita (Italian) | 2.7 | | deu (German) | 6.9 | jpn (Japanese) | 2.6 | | spa (Spanish) | 6.0 | pol (Polish) | 1.7 | | fra (French) | 4.8 | por (Portuguese) | 1.7 | | rus (Russian) | 3.8 | tur (Turkish) | 1.5 | | nld (Dutch) | 2.8 | Other | 13.0 | > **Note**: Languages are specified through their corresponding ISO 639-3 code (e.g., "eng", "deu", "fra"). ### Topics | | Topic | % | |------------------|------------------------------------------------------|------| | ✈️ | Traveling and Hospitality | 34.1 | | 🛒 | Products and Commercial Services | 19.8 | | ❤️ | Healthcare Services, Wellness, and Lifestyle | 13.0 | | 🎵 | Entertainment, Recreation, and Leisure | 9.7 | | 🎓 | Employment, Education, and Training | 9.5 | | 🏦 | Banking, Financial Services, and Insurance | 6.0 | | ⚖️ | Legal Services, Regulations, and Government | 4.0 | | ℹ️ | General Information and Other | 3.9 | ### Question Types | Question Type | % | |------------------------------|------| | What | 24.3 | | How | 17.1 | | Is, are, do, does | 16.7 | | Which | 11.4 | | Can, will, may, shall | 7.3 | | [No Question Word] | 6.5 | | When | 6.0 | | Where | 5.5 | | Why | 3.0 | | Who | 2.0 | ## Structure The raw Q&A dataset is offered in 75 language-specific subsets with one default split. Within these subsets, each row represents a single QA pair. The primary fields include: - **`id`**: Unique identifier for each sample. - **`origin`**: Specifies the FAQ's website through scheme, host, and optionally, port. - **`url`**: URL of the webpage where the QA pair was extracted. - **`question`**: Natural-language question. - **`answer`**: Corresponding answer text. - *Optionally:* **`topic`**: One of eight broad categories (e.g., “Traveling and Hospitality”). - *Optionally:* **`question_type`**: (What, Who, How, Why, When, Where, etc.). > **Note**: The Q&A dataset does not include an official train/validation/test split. If you require such splits (e.g., for training models), you can create them programmatically or consider using the separate [WebFAQ retrieval dataset splits](#overview). Furthermore, the language is not given as a field in the dataset, as it is specified through the selected subset. The topic and question type are optional fields that are only present in the 49 language subsets with ≥100 websites. ## Examples Below is a sample record in JSON-like format: ```json { "id": "256a247bc6fa1f366512affe9327cc85", "origin": "http://0.suzhoulvsen.com", "url": "http://0.suzhoulvsen.com/pimcore", "question": "What is Pimcore Development?", "answer": "Skynet Technologies uses Pimcore as a data experience management platform to create websites, custom applications, PIM solutions, DAM solutions, ecommerce websites, and more.", "topic": "Employment, Education and Training", "question_type": "What" } ``` Below is an example in Python using the 🤗 [Datasets library](https://github.com/huggingface/datasets): ```python from datasets import load_dataset dataset = load_dataset("PaDaS-Lab/webfaq", "eng")["default"] print(dataset[0]) # Example output: # { # 'id': '256a247bc6fa1f366512affe9327cc85', # 'origin': 'http://0.suzhoulvsen.com', # 'url': 'http://0.suzhoulvsen.com/pimcore', # 'question': 'What is Pimcore Development?', # 'answer': 'Skynet Technologies uses Pimcore as a data experience management platform to create websites, custom applications, PIM solutions, DAM solutions, ecommerce websites, and more.', # 'topic': 'Employment, Education and Training', # 'question_type': 'What' # } ``` ## Considerations Please note the following considerations when using the collected QAs: - *[Q&A Dataset]* **Risk of Duplicate or Near-Duplicate Content**: The raw Q&A dataset is large and includes minor paraphrases. - *[Retrieval Dataset]* **Sparse Relevance**: As raw FAQ data, each question typically has one “best” (on-page) answer. Additional valid answers may exist on other websites but are not labeled as relevant. - **Language Detection Limitations**: Some QA pairs mix languages, or contain brand names, which can confuse automatic language classification. - **No Guarantee of Factual Accuracy**: Answers reflect the content of the source websites. They may include outdated, biased, or incorrect information. - **Copyright and Privacy**: Please ensure compliance with any applicable laws and the source website’s terms. ## License The **Collection of WebFAQ Datasets** is shared under [**Creative Commons Attribution 4.0 (CC BY 4.0)**](https://creativecommons.org/licenses/by/4.0/) license. > **Note**: The dataset is derived from public webpages in Common Crawl snapshots (2022–2024) and intended for **research purposes**. Each FAQ’s text is published by the original website under their terms. Downstream users should verify any usage constraints from the **original websites** as well as [Common Crawl’s Terms of Use](https://commoncrawl.org/terms-of-use/). ## Citation If you use this dataset in your research, please consider citing the associated paper: ```bibtex @misc{dinzinger2025webfaq, title={WebFAQ: A Multilingual Collection of Natural Q&A Datasets for Dense Retrieval}, author={Michael Dinzinger and Laura Caspari and Kanishka Ghosh Dastidar and Jelena Mitrović and Michael Granitzer}, year={2025}, eprint={2502.20936}, archivePrefix={arXiv}, primaryClass={cs.CL} } ``` ## Contact For inquiries and feedback, please feel free to contact us via E-Mail ([michael.dinzinger@uni-passau.de](mailto:michael.dinzinger@uni-passau.de)) or start a discussion on HuggingFace or GitHub. ## Acknowledgement We thank the Common Crawl and Web Data Commons teams for providing the underlying data, and all contributors who helped shape the WebFAQ project. ### Thank you We hope the **Collection of WebFAQ Datasets** serves as a valuable resource for your research. Please consider citing it in any publications or projects that use it. If you encounter issues or want to contribute improvements, feel free to get in touch with us on HuggingFace or GitHub. Happy researching!