Can we trust LLM Self-Explanations for Entity Resolution?
Summary: Large-scale evaluation across LLMs, datasets, and prompts finds ER self-explanations unstable, unfaithful, and poorly counterfactually aligned. ELLMER uses them as priors for cheaper post-hoc explanations, matching quality at up to 10× lower cost. (summarized by gpt-5.6-luna on Aug 28 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Tommaso Teofili (Elastic; Roma Tre University)
- 2. Donatella Firmani (Sapienza University)
- 3. Nick Koudas (University of Toronto)
- 4. Paolo Merialdo (Roma Tre University)
- 5. Divesh Srivastava (AT&T)
BibTeX Citation
@article{teofili_vldb26,
title = {{Can we trust LLM Self-Explanations for Entity Resolution?}},
author = {Teofili, Tommaso and Firmani, Donatella and Koudas, Nick and Merialdo, Paolo and Srivastava, Divesh},
journal = {PVLDB},
series = {{VLDB} '26},
volume = {19},
number = {11},
pages = {3538--3551},
doi = {10.14778/3836663.3836707},
url = {https://doi.org/10.14778/3836663.3836707},
year = {2026}
}Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 8 of 8 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 134 | Deep Entity Matching with Pre-Trained Language Models | 2021 | VLDB | 0.00030043481 |
| 158 | Deep Learning for Entity Matching: A Design Space Exploration | 2018 | SIGMOD | 0.00028046388 |
| 4,033 | Dual-Objective Fine-Tuning of BERT for Entity Matching | 2021 | VLDB | 6.8394738e-05 |
| 5,149 | LLM for Data Management | 2024 | VLDB | 6.2551644e-05 |
| 5,540 | Pre-trained Embeddings for Entity Resolution: An Experimental Analysis | 2023 | VLDB | 6.0907591e-05 |
| 7,401 | BrewER: Entity Resolution On-Demand | 2023 | VLDB | 5.5359939e-05 |
| 8,888 | Optimized Batch Prompting for Cost-effective LLMs | 2025 | VLDB | 5.2559789e-05 |
| 11,906 | CERTEM: Explaining and Debugging Black-box Entity Resolution Systems with CERTA | 2022 | VLDB | 4.9793485e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 10,462 | Generalized Entity Matching with Adaptivity via Large Language Models | 2026 | SIGMOD |
| 2 | 5,946 | Exploiting Context Analysis for Combining Multiple Entity Resolution Systems | 2009 | SIGMOD |
| 3 | 10,841 | Bolt-on, Verifiable Provenance for LLM-Powered Data Processing | 2026 | VLDB |
| 4 | 7,516 | SystemER: A Human-in-the-loop System for Explainable Entity Resolution | 2019 | VLDB |
| 5 | 7,012 | Reliable Text-to-SQL with Adaptive Abstention | 2025 | SIGMOD |
| 6 | 8,316 | ThriftLLM: On Cost-Effective Selection of Large Language Models for Classification Queries | 2025 | VLDB |
| 7 | 3,695 | Revisiting Prompt Engineering via Declarative Crowdsourcing | 2024 | CIDR |
| 8 | 5,540 | Pre-trained Embeddings for Entity Resolution: An Experimental Analysis | 2023 | VLDB |
| 9 | 10,527 | In-context Clustering-based Entity Resolution with Large Language Models: A Design Space Exploration | 2026 | SIGMOD |
| 10 | 11,004 | LLM-CER: An Interactive System for In-context Clustering-based Entity Resolution with Large Language Models | 2026 | VLDB |