Discerning and Resolving Knowledge Conflicts through Adaptive Decoding with Contextual Information-Entropy Constraint
This work addresses a critical issue for users of large language models in realistic applications where external knowledge integration is necessary, representing an incremental improvement over existing specialized decoding methods.
The paper tackles the problem of knowledge conflicts in large language models, where external contextual knowledge clashes with internal parametric knowledge, by proposing an adaptive decoding method called COIECD that discerns and resolves conflicts, resulting in improved faithfulness to conflicting contexts and maintained high performance in non-conflict scenarios.
Large language models 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 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- Our experiments show that COIECD exhibits strong performance and robustness over knowledge conflicts in realistic datasets. Code is available.