Investigating Tool-Memory Conflicts in Tool-Augmented LLMs

Jiali Cheng, Rui Pan, Hadi Amiri

Keywords: Knowledge Conflict, LLM

Abstract:

Tool-augmented large language models (LLMs)
have powered many applications. However, they
are likely to suffer from knowledge conflict. In
this paper, we propose a new type of knowledge conflict – Tool-Memory Conflict (TMC),
where the internal parametric knowledge contradicts with the external tool knowledge for toolaugmented LLMs. We find that existing LLMs,
though powerful, suffer from TMC, especially
on STEM-related tasks. We also uncover that
under different conditions, tool knowledge and
parametric knowledge may be prioritized differently. We then evaluate existing conflict resolving
techniques, including prompting-based and RAGbased methods. Results show that none of these
approaches can effectively resolve tool-memory
conflicts.

Submission Number: 167