WebFAQ 2.0: A Multilingual QA Dataset with Mined Hard Negatives for Dense Retrieval

TL;DR

WebFAQ 2.0 is a large, multilingual FAQ dataset with mined hard negatives designed to improve dense retrieval models, enabling more diverse and effective training for cross-lingual question answering.

Contribution

It introduces the largest multilingual FAQ dataset with a novel data collection method and a hard negatives dataset for training dense retrievers, advancing multilingual IR research.

Findings

Enhanced multilingual coverage with 198M QA pairs

Effective training of dense retrievers using mined hard negatives

Openly available datasets and scripts for community use

Abstract

We introduce WebFAQ 2.0, a new version of the WebFAQ dataset, containing 198 million FAQ-based natural question-answer pairs across 108 languages. Compared to the previous version, it significantly expands multilingual coverage and the number of bilingual aligned QA pairs to over 14.3M, making it the largest FAQ-based resource. Unlike the original release, WebFAQ 2.0 uses a novel data collection strategy that directly crawls and extracts relevant web content, resulting in a substantially more diverse and multilingual dataset with richer context through page titles and descriptions. In response to community feedback, we also release a hard negatives dataset for training dense retrievers, with 1.25M queries across 20 languages. These hard negatives were mined using a two-stage retrieval pipeline and include cross-encoder scores for 200 negatives per query. We further show how this resource enables two primary fine-tuning strategies for dense retrievers: Contrastive Learning with MultipleNegativesRanking loss, and Knowledge Distillation with MarginMSE loss. WebFAQ 2.0 is not a static resource but part of a long-term effort. Since late 2025, structured FAQs are being regularly released through the Open Web Index, enabling continuous expansion and refinement. We publish the datasets and training scripts to facilitate further research in multilingual and cross-lingual IR. The dataset itself and all related resources are publicly available on GitHub and HuggingFace.