Benchmarking Knowledge Editing using Logical Rules
T. Moteu Ngoli, N.J. Kouagou, H.M.A. Zahera, A.-C. Ngonga Ngomo, in: Proceedings of the 24th International Semantic Web Conference (ISWC 2025), n.d.
Download
No fulltext has been uploaded.
Conference Paper
| Accepted
| English
Author
Department
Abstract
Large Language Models (LLMs) are increasingly deployed in real-world applications that require access to up-to-date knowledge. However, retraining LLMs is computationally expensive. Therefore, knowledge editing techniques are crucial for maintaining current information and correcting erroneous assertions within pre-trained models. Current benchmarks for knowledge editing primarily focus on recalling edited facts, often neglecting their logical consequences. To address this limitation, we introduce a new benchmark designed to evaluate how knowledge editing methods handle the logical consequences of a single fact edit. Our benchmark extracts relevant logical rules from a knowledge graph for a given edit. Then, it generates multi-hop questions based on these rules to assess the impact on logical consequences. Our findings indicate that while existing knowledge editing approaches can accurately insert direct assertions into LLMs, they frequently fail to inject entailed knowledge. Specifically, experiments with popular methods like ROME and FT reveal a substantial performance gap, up to 24%, between evaluations on directly edited knowledge and on entailed knowledge. This highlights the critical need for semantics-aware evaluation frameworks in knowledge editing.
Publishing Year
Proceedings Title
Proceedings of the 24th International Semantic Web Conference (ISWC 2025)
Conference
The 24th International Semantic Web Conference (ISWC 2025)
Conference Location
Nara, Japan
Conference Date
2025.11.2 – 2025.11.6
LibreCat-ID
Cite this
Moteu Ngoli T, Kouagou NJ, Zahera HMA, Ngonga Ngomo A-C. Benchmarking Knowledge Editing using Logical Rules. In: Proceedings of the 24th International Semantic Web Conference (ISWC 2025).
Moteu Ngoli, T., Kouagou, N. J., Zahera, H. M. A., & Ngonga Ngomo, A.-C. (n.d.). Benchmarking Knowledge Editing using Logical Rules. Proceedings of the 24th International Semantic Web Conference (ISWC 2025). The 24th International Semantic Web Conference (ISWC 2025), Nara, Japan.
@inproceedings{Moteu Ngoli_Kouagou_Zahera_Ngonga Ngomo, title={Benchmarking Knowledge Editing using Logical Rules}, booktitle={Proceedings of the 24th International Semantic Web Conference (ISWC 2025)}, author={Moteu Ngoli, Tatiana and Kouagou, N’Dah Jean and Zahera, Hamada Mohamed Abdelsamee and Ngonga Ngomo, Axel-Cyrille} }
Moteu Ngoli, Tatiana, N’Dah Jean Kouagou, Hamada Mohamed Abdelsamee Zahera, and Axel-Cyrille Ngonga Ngomo. “Benchmarking Knowledge Editing Using Logical Rules.” In Proceedings of the 24th International Semantic Web Conference (ISWC 2025), n.d.
T. Moteu Ngoli, N. J. Kouagou, H. M. A. Zahera, and A.-C. Ngonga Ngomo, “Benchmarking Knowledge Editing using Logical Rules,” presented at the The 24th International Semantic Web Conference (ISWC 2025), Nara, Japan.
Moteu Ngoli, Tatiana, et al. “Benchmarking Knowledge Editing Using Logical Rules.” Proceedings of the 24th International Semantic Web Conference (ISWC 2025).