Apr 23
The Tool-Overuse Illusion: Why Does LLM Prefer External Tools over Internal Knowledge?
★★★★★
significance 3/5
Researchers identify a phenomenon called 'tool-overuse' where LLMs unnecessarily use external tools instead of relying on internal knowledge. The paper proposes a knowledge-aware alignment strategy and balanced reward signals to reduce unnecessary tool calls while improving accuracy.
Why it matters
Unnecessary tool dependency threatens operational efficiency and highlights a fundamental misalignment in how models weigh internal knowledge against external APIs.
Tags
#llm #tool-use #reasoning #optimization #epistemic illusionRelated coverage
- Global South OpportunitiesPivotal Research Fellowship 2026 (Q3): AI Safety Research Opportunity - Global South Opportunities
- arXiv cs.AIAn Intelligent Fault Diagnosis Method for General Aviation Aircraft Based on Multi-Fidelity Digital Twin and FMEA Knowledge Enhancement
- arXiv cs.AIPExA: Parallel Exploration Agent for Complex Text-to-SQL
- arXiv cs.AIThe Power of Power Law: Asymmetry Enables Compositional Reasoning
- arXiv cs.AIOn the Existence of an Inverse Solution for Preference-Based Reductions in Argumentation