# Injecting the BM25 Score as Text Improves BERT-Based Re-rankers

arXiv: https://arxiv.org/abs/2301.09728 · HTML: https://mlsanity.com/paper/2301.09728/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Jan 2023.

## Built on by (most cited first, 17 of 17)

- [Knowledge Graph-Guided Retrieval Augmented Generation](https://mlsanity.com/paper/2502.06864/) (2025)
- [FinSage: A Multi-aspect RAG System for Financial Filings Question Answering](https://mlsanity.com/paper/2504.14493/) (2025)
- [RAM: Towards an Ever-Improving Memory System by Learning from Communications](https://mlsanity.com/paper/2404.12045/) (2024)
- [InsertRank: LLMs can reason over BM25 scores to Improve Listwise Reranking](https://mlsanity.com/paper/2506.14086/) (2025)
- [TeleOracle: Fine-Tuned Retrieval-Augmented Generation with Long-Context Support for Network](https://mlsanity.com/paper/2411.02617/) (2024)
- [Enhancing Documents with Multidimensional Relevance Statements in Cross-encoder Re-ranking](https://mlsanity.com/paper/2306.10979/) (2023)
- [Identification of Regulatory Requirements Relevant to Business Processes: A Comparative Study on Generative AI, Embedding-based Ranking, Crowd and Expert-driven Methods](https://mlsanity.com/paper/2401.02986/) (2024)
- [Semantic Captioning: Benchmark Dataset and Graph-Aware Few-Shot In-Context Learning for SQL2Text](https://mlsanity.com/paper/2501.03166/) (2025)
- [Memory-Augmented Architecture for Long-Term Context Handling in Large Language Models](https://mlsanity.com/paper/2506.18271/) (2025)
- [Hawk: Harnessing Hardware-Aware Knowledge for High-Performance NPU Kernel Generation](https://mlsanity.com/paper/2607.01590/) (2026)
- [CIKMar: A Dual-Encoder Approach to Prompt-Based Reranking in Educational Dialogue Systems](https://mlsanity.com/paper/2408.08805/) (2024)
- [MST-R: Multi-Stage Tuning for Retrieval Systems and Metric Evaluation](https://mlsanity.com/paper/2412.10313/) (2024)
- [Scalable Unit Harmonization in Medical Informatics via Bayesian-Optimized Retrieval and Transformer-Based Re-ranking](https://mlsanity.com/paper/2505.00810/) (2025)
- [REGENT: Relevance-Guided Attention for Entity-Aware Multi-Vector Neural Re-Ranking](https://mlsanity.com/paper/2510.11592/) (2025)
- [A Multi-agent Text2SQL Framework using Small Language Models and Execution Feedback](https://mlsanity.com/paper/2512.18622/) (2025)
- [Retrieval-Enhanced Real Estate Appraisal](https://mlsanity.com/paper/2603.12986/) (2026)
- [Embedding Surgery: Localized Updates for Adaptive Ranking Correction in Dense Retrieval](https://mlsanity.com/paper/2609.05110/) (2026)

## Abstract (arXiv)

In this paper we propose a novel approach for combining first-stage lexical retrieval models and Transformer-based re-rankers: we inject the relevance score of the lexical model as a token in the middle of the input of the cross-encoder re-ranker. It was shown in prior work that interpolation between the relevance score of lexical and BERT-based re-rankers may not consistently result in higher effectiveness. Our idea is motivated by the finding that BERT models can capture numeric information. We compare several representations of the BM25 score and inject them as text in the input of four different cross-encoders. We additionally analyze the effect for different query types, and investigate the effectiveness of our method for capturing exact matching relevance. Evaluation on the MSMARCO Passage collection and the TREC DL collections shows that the proposed method significantly improves over all cross-encoder re-rankers as well as the common interpolation methods. We show that the improvement is consistent for all query types. We also find an improvement in exact matching capabilities over both BM25 and the cross-encoders. Our findings indicate that cross-encoder re-rankers can efficiently be improved without additional computational burden and extra steps in the pipeline by explicitly adding the output of the first-stage ranker to the model input, and this effect is robust for different models and query types.