{"id":"rifts","title":"Navigating Rifts in Human-LLM Grounding: Study and Benchmark","authors":[{"given":"Omar","family":"Shaikh"},{"given":"Hussein","family":"Mozannar"},{"given":"Gagan","family":"Bansal"},{"given":"Adam","family":"Fourney"},{"given":"Eric","family":"Horvitz"}],"first_public":"2025-03-18","date_note":"First arXiv version.","date_source":"https://arxiv.org/abs/2503.13975","year":2025,"venue":"ACL 2025","paper_url":"https://aclanthology.org/2025.acl-long.1016/","arxiv":"2503.13975","category":"evaluation","relationship":"uses","dataset_version":"Non-toxic English WildChat conversations, sampled at one conversation per user for the main grounding analysis; RIFTS selects initial prompts from WildChat.","summary":"Analyzes how users and assistants establish mutual understanding, alongside Bing Chat and human-human MultiWOZ conversations. It releases grounding annotations, a forecasting approach, and RIFTS to test when assistants should clarify, ask follow-up questions, or proceed.","finding":"RIFTS contains 1,740 tasks. The evaluated assistants rarely take useful grounding initiative; forecasting-based prompting improves their performance.","evidence_url":"https://arxiv.org/html/2503.13975v2","evidence_note":"Section 4: data filters. Section 6 and Appendix A: benchmark construction and evaluation.","bibtex_key":"rifts2025","record_url":"https://yuntiandeng.com/wildchat/papers/rifts/","relationship_label":"Uses WildChat","citation_title":"Navigating Rifts in Human-LLM Grounding: Study and Benchmark","citation_type":"paper-conference","bibtex_type":"inproceedings","citation_url":"https://aclanthology.org/2025.acl-long.1016/","citation_year":2025,"citation_note":"ACL 2025 proceedings paper.","doi":"10.18653/v1/2025.acl-long.1016","code_url":"https://github.com/microsoft/rifts","container_title":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","pages":"20832–20847","publisher":"Association for Computational Linguistics"}