# Beyond Scale and Generation: Understanding Language Model-based Entity Matching

arXiv: https://arxiv.org/abs/2607.24688 · HTML: https://mlsanity.com/paper/2607.24688/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 69 references that are in the map.
- First posted on arXiv 27 Jul 2026.

## Builds on (most cited first, 28 of 28)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [Deep Entity Matching with Pre-Trained Language Models](https://mlsanity.com/paper/2004.00584/) (2020)
- [Rethinking LLM Memorization through the Lens of Adversarial Compression](https://mlsanity.com/paper/2404.15146/) (2024)
- [DeepER -- Deep Entity Resolution](https://mlsanity.com/paper/1710.00597/) (2017)
- [Low-resource Deep Entity Resolution with Transfer and Active Learning](https://mlsanity.com/paper/1906.08042/) (2019)
- [Neural Networks for Entity Matching: A Survey](https://mlsanity.com/paper/2010.11075/) (2020)
- [Match, Compare, or Select? An Investigation of Large Language Models for Entity Matching](https://mlsanity.com/paper/2405.16884/) (2024)
- [Supervised Contrastive Learning for Product Matching](https://mlsanity.com/paper/2202.02098/) (2022)
- [Sudowoodo: Contrastive Self-supervised Learning for Multi-purpose Data Integration and Preparation](https://mlsanity.com/paper/2207.04122/) (2022)
- [WDC Products: A Multi-Dimensional Entity Matching Benchmark](https://mlsanity.com/paper/2301.09521/) (2023)
- [Machamp: A Generalized Entity Matching Benchmark](https://mlsanity.com/paper/2106.08455/) (2021)
- [Fine-tuning Large Language Models for Entity Matching](https://mlsanity.com/paper/2409.08185/) (2024)
- [Bridging the Gap between Reality and Ideality of Entity Matching: A Revisiting and Benchmark Re-Construction](https://mlsanity.com/paper/2205.05889/) (2022)
- [DAME: Domain Adaptation for Matching Entities](https://mlsanity.com/paper/2204.09244/) (2022)
- [Through the Fairness Lens: Experimental Analysis and Evaluation of Entity Matching](https://mlsanity.com/paper/2307.02726/) (2023)
- [Deep Transfer Learning for Multi-source Entity Linkage via Domain Adaptation](https://mlsanity.com/paper/2110.14509/) (2021)
- [SC-Block: Supervised Contrastive Blocking within Entity Resolution Pipelines](https://mlsanity.com/paper/2303.03132/) (2023)
- [Ground Truth Inference for Weakly Supervised Entity Matching](https://mlsanity.com/paper/2211.06975/) (2022)
- [Contrastive Entity Coreference and Disambiguation for Historical Texts](https://mlsanity.com/paper/2406.15576/) (2024)
- [BEACON: Budget-Aware Entity Matching Across Domains (Extended Technical Report)](https://mlsanity.com/paper/2603.11391/) (2026)
- [Heterogeneity in Entity Matching: A Survey and Experimental Analysis](https://mlsanity.com/paper/2508.08076/) (2025)
- [Can we trust LLM Self-Explanations for Entity Resolution?](https://mlsanity.com/paper/2606.01210/) (2026)
- [Entity Resolution via Batched Oracle Queries](https://mlsanity.com/paper/2606.24407/) (2026)
- [Understanding Domain-Aware Distribution Alignment in Budgeted Entity Matching](https://mlsanity.com/paper/2606.27342/) (2026)

## Abstract (arXiv)

Entity matching identifies records that refer to the same real-world entity. Language models can be adapted to this task through bi-encoder, cross-encoder, and generative matcher architectures. However, prior studies often conflate matcher architecture with differences in model backbone, model variant(reflecting different pretraining objectives), and model size, making it difficult to isolate the sources of performance gains. We address this issue through a controlled factorial study spanning three matcher architectures, three model variants and three model sizes from the Qwen3 family, and nine datasets, totaling 1,215 fine-tuning runs. We also evaluate cross-dataset transferability and computational cost. Our results show that model variant is critical for bi-encoders: embedding-oriented variants provide stronger initialization and more favorable representation geometry predictive of downstream matching performance. Cross-encoders retain a consistent advantage over bi-encoders because they jointly encode record pairs rather than representing each record independently, although larger models partially narrow this gap. Generative matchers do not universally outperform cross-encoders. Instead, their advantages concentrate under distribution shift, including subtle unseen differences in record schemas and cross-dataset transfer. We further find that larger models rely more heavily on shortcut learning and therefore do not necessarily perform better. These findings clarify the factors underlying performance differences across matcher architectures and motivate future research and benchmark designs that better disentangle architectural choices from model-level factors while explicitly evaluating distribution shift and cross-dataset transferability. We release our experimental results, code, training scripts, and evaluation data at https://github.com/Jantory/llm-trained-matcher.