chengpingan/ExpandR_retriever_training_data
Viewer • Updated • 709k • 52 • 1
This checkpoint is the document retrieval component of ExpandR. It starts from AnchorDR and learns from expanded queries, positive documents and negative documents through supervised contrastive training. The paired query expansion generator learns from retrieval feedback using DPO. The paper evaluates the two components together on BEIR. Evaluation script
Training data · Training and evaluation code
The first preprint appeared as LLM-QE: Improving Query Expansion by Aligning Large Language Models with Ranking Preferences. The archival citation is Sijia Yao, Pengcheng Huang, Zhenghao Liu, Yu Gu, Yukun Yan, Shi Yu, and Ge Yu (EMNLP 2025), ExpandR: Teaching Dense Retrievers Beyond Queries with LLM Guidance. BibTeX
Base model
yiqingx/AnchorDR