paper-with-me

홈 › Papers

The Tool-Overuse Illusion: Why Does LLM Prefer External Tools over Internal Knowledge?

2026-03-03 · Yirong Zeng, Shen You, Yufei Liu, Qunyao Du, Xiao Ding, Yutai Hou, Yuxian Wang, Wu Ning, Haonan Song, Dandan Tu, Bibo Cai, Ting Liu arxiv

Equipping LLMs with external tools effectively addresses internal reasoning limitations. However, it introduces a critical yet under-explored phenomenon: tool overuse, the unnecessary tool-use during reasoning. In this paper, we first reveal this phenomenon is pervasive across diverse LLMs. We then experimentally elucidate its underlying mechanisms through two key lenses: (1) First, by analyzing tool-use behavior across different internal knowledge availability regions, we identify a \textit{knowledge epistemic illusion}: models misjudge internal knowledge boundaries and fail to accurately perceive their actual knowledge availability. To mitigate this, we propose a knowledge-aware epistemic boundary alignment strategy based on direct preference optimization, which reduces tool usage in by 82.8\% while yielding an accuracy improvement. (2) Second, we establish a causal link between reward structures and tool-use behavior by visualizing the tool-augmented training process. It reveals that \textit{outcome-only rewards} inadvertently encourage tool overuse by rewarding only final correctness, regardless of tool efficiency. To verify this, we balance reward signals during training rather than relying on outcome-only rewards, cutting unnecessary tool calls by 66.7\% (7B) and 60.7\% (32B) without sacrificing accuracy. Finally, we provide theoretical justification in this two lenses to understand tool overuse.

📄 PDF Abstract BibTeX arXiv:2604.19749

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

SMART: Self-Aware Agent for Tool Overuse Mitigation

2025-02-17 · Cheng Qian, Emre Can Acikgoz, Hongru Wang, Xiusi Chen 외

Current Large Language Model (LLM) agents demonstrate strong reasoning and tool use capabilities, but often lack self-awareness, failing to balance these approaches effectively. This imbalance leads to Tool Overuse, wher…

GSM8KLarge Language Model

Word Overuse and Alignment in Large Language Models: The Influence of Learning from Human Feedback

2025-08-03 · Tom S. Juzek, Zina B. Ward arxiv

Large Language Models (LLMs) are known to overuse certain terms like "delve" and "intricate." The exact reasons for these lexical choices, however, have been unclear. Using Meta's Llama model, this study investigates the…

Reinforcement Learning

From Knowing to Acting: Benchmarking Self-Awareness Capability of LLM Agents

2026-06-09 · Yifan Li, Shengbin Yue, Boyu Feng, Jinhu Qi 외 arxiv

The integration of external tools has transitioned LLM agents from passive responders to autonomous systems. However, current benchmarks prioritize execution success, neglecting self-awareness capability, the ability to …

Learning When Not to Act: Mitigating Tool Abuse in Agentic Reinforcement Learning

2026-06-01 · Liuji Chen, Dianxing Tang, Xing Shi, Dingshuo Chen 외 arxiv

Agentic reinforcement learning can induce tool abuse, where models overuse external tools even for queries solvable by internal reasoning. Existing approaches mitigate this issue with uniform tool-use penalties or hard l…

Reinforcement Learning

Illusions in Humans and AI: How Visual Perception Aligns and Diverges

2025-08-17 · Jianyi Yang, Junyi Ye, Ankan Dash, Guiling Wang arxiv

By comparing biological and artificial perception through the lens of illusions, we highlight critical differences in how each system constructs visual reality. Understanding these divergences can inform the development …