{"id":"wildchat-aqa","title":"From Chat Logs to Collective Insights: Aggregative Question Answering","authors":[{"given":"Wentao","family":"Zhang"},{"given":"Woojeong","family":"Kim"},{"given":"Yuntian","family":"Deng"}],"first_public":"2025-05-29","date_note":"First arXiv version.","date_source":"https://arxiv.org/abs/2505.23765","year":2025,"venue":"EMNLP 2025","paper_url":"https://aclanthology.org/2025.emnlp-main.1667/","arxiv":"2505.23765","category":"evaluation","relationship":"uses","dataset_version":"A filtered WildChat-Full release of roughly 990,000 conversations, reduced to 182,330 conversations after deduplication, length limits, and an active-user requirement.","summary":"Introduces aggregative question answering: answering questions that require combining evidence across many conversations. WildChat-AQA pairs conversation text and metadata with inferred topics and keywords, and compares retrieval, fine-tuning, and the PROBE retrieval method.","finding":"The benchmark contains 6,027 questions over 182,330 conversations. PROBE improves over standard retrieval, with retrieval and reasoning errors remaining in the evaluated systems.","evidence_url":"https://arxiv.org/html/2505.23765v2","evidence_note":"Section 2 and Appendix E: construction, model-generated questions, and human quality checks. Section 4: ranking quality, retrieval, and computational cost.","bibtex_key":"wildchataqa2025","record_url":"https://yuntiandeng.com/wildchat/papers/wildchat-aqa/","relationship_label":"Uses WildChat","citation_title":"From Chat Logs to Collective Insights: Aggregative Question Answering","citation_type":"paper-conference","bibtex_type":"inproceedings","citation_url":"https://aclanthology.org/2025.emnlp-main.1667/","citation_year":2025,"citation_note":"EMNLP 2025 proceedings paper.","doi":"10.18653/v1/2025.emnlp-main.1667","code_url":"https://github.com/yuntian-group/wildchat_aggregative_question_answering","container_title":"Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing","pages":"32811–32850","publisher":"Association for Computational Linguistics"}