Discerning and Resolving Knowledge Conflicts through Adaptive Decoding with Contextual Information-Entropy Constraint

Xiaowei Yuan, Zhao Yang, Yequan Wang, Shengping Liu, Jun Zhao, Kang Liu


Abstract
Large language models (LLMs) internalize enormous parametric knowledge during pre-training. Concurrently, realistic applications necessitate external contextual knowledge to aid models on the underlying tasks. This raises a crucial dilemma known as knowledge conflicts, where the contextual knowledge clashes with the parametric knowledge. However, existing decoding works are specialized in resolving knowledge conflicts and could inadvertently deteriorate performance in absence of conflicts. In this paper, we propose an adaptive decoding method, termed as contextual information-entropy constraint decoding (COIECD), to discern whether the knowledge conflicts occur and resolve them. It can improve the model’s faithfulness to conflicting context, and simultaneously maintain high performance among non-conflicting context. Our experiments show that COIECD exhibits strong performance and robustness over knowledge conflicts in realistic datasets.
Anthology ID:
2024.findings-acl.234
Volume:
Findings of the Association for Computational Linguistics: ACL 2024
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
3903–3922
Language:
URL:
https://aclanthology.org/2024.findings-acl.234
DOI:
10.18653/v1/2024.findings-acl.234
Bibkey:
Cite (ACL):
Xiaowei Yuan, Zhao Yang, Yequan Wang, Shengping Liu, Jun Zhao, and Kang Liu. 2024. Discerning and Resolving Knowledge Conflicts through Adaptive Decoding with Contextual Information-Entropy Constraint. In Findings of the Association for Computational Linguistics: ACL 2024, pages 3903–3922, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
Discerning and Resolving Knowledge Conflicts through Adaptive Decoding with Contextual Information-Entropy Constraint (Yuan et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-acl.234.pdf