Impact of WebFAQ 2.0 Data Collection on Multilingual Dense Retriever Generalization

We present WebFAQ, a large-scale collection of open-domain question answering datasets derived from FAQ-style schema.org annotations. In total, the data collection consists of 96 million natural question-answer (QA) pairs across 75 languages, including 47 million (49\%) non-English samples. WebFAQ further serves as the foundation for 20 monolingual retrieval benchmarks with a total size of 11.2 million QA pairs (5.9 million non-English). These datasets are carefully curated through refined filtering and near-duplicate detection, yielding high-quality resources for training and evaluating multil

Research goal: What is the impact of the novel web-crawled data collection strategy in WebFAQ 2.0 on the domain generalization capabilities of multilingual dense retrievers when evaluated on out-of-domain QA datasets across diverse language families?

Autonomous synthesis report generated by SOVEREIGN Research Kernel. Tribunal consensus score: 9.2/10.