RankT5: Fine-Tuning T5 for Text Ranking with Ranking Losses

By Honglei Zhuang, Zhen Qin, R. Jagerman, Kai Hui, Ji Ma, Jing Lu, Jianmo Ni, Xuanhui Wang, Michael Bendersky. 2022

Subfields: Information Retrieval & Search, Natural Language Processing

Citations in the corpus

• CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking (2024, unknown): “Furthermore, Sun et al. (2023) show that instruction-tuned LLMs can outperform traditional supervised cross-encoders (Nogueira et al., 2020; Zhuang et al., 2023b) in zero-shot reranking settings.”
• A Survey of Model Architectures in Information Retrieval (2025, unknown): “Zhuang et al. (2023a) proposed to fine-tune T5 to produce a numerical relevance score, and optimize the model with ranking losses like R ANK N ET (Burges, 2010).”
• SweRank+: Multilingual, Multi-Turn Code Ranking for Software Issue Localization (2025, unknown): “This approach leverages the instruction-following capabilities of large language models and has proven more effective than pointwise (Zhuang et al., 2023, 2024) or pairwise ranking (Qin et al., 2024).”
• LLM-confidence reranker: A training-free approach for enhancing retrieval-augmented generation systems (2026, unknown): “• RankT5 (Zhuang et al., 2023): A Transformer-based sequence-to-sequence model designed for reranking tasks.”