When to Trust Tools? Adaptive Tool Trust Calibration For Tool-Integrated Math Reasoning
2026-04-09 • Computation and Language
Computation and Language
AI summaryⓘ
The authors studied how big language models use tools during reasoning tasks but found these models often ignore tool results even when the tools are correct. They call this problem "Tool Ignored," meaning the model doesn’t know when to trust the tool’s answers. To fix this, the authors developed Adaptive Tool Trust Calibration (ATTC), which helps the model decide when to trust or ignore tool outputs based on confidence scores. Tests showed ATTC improved model performance by reducing ignored correct tool results.
Large reasoning modelsTool-Integrated ReasoningTool IgnoredAdaptive Tool Trust Calibrationconfidence scoremodel trustcode executionreasoning trajectory
Authors
Ruotao Xu, Yixin Ji, Yu Luo, Jinpeng Li, Dong Li, Peifeng Li, Juntao Li, Min Zhang
Abstract
Large reasoning models (LRMs) have achieved strong performance enhancement through scaling test time computation, but due to the inherent limitations of the underlying language models, they still have shortcomings in tasks that require precise computation and extensive knowledge reserves. Tool-Integrated Reasoning (TIR) has emerged as a promising paradigm that incorporates tool call and execution within the reasoning trajectory. Although recent works have released some powerful open-source TIR models, our analysis reveals that these models still suffer from critical deficiencies. We find that when the reasoning of the model conflicts with the tool results, the model tends to believe in its own reasoning. And there are cases where the tool results are correct but are ignored by the model, resulting in incorrect answers, which we define as "Tool Ignored''. This indicates that the model does not know when to trust or ignore the tool. To overcome these limitations, We introduce Adaptive Tool Trust Calibration (ATTC), a novel framework that guides the model to adaptively choose to trust or ignore the tool results based on the confidence score of generated code blocks. The experimental results from various open-source TIR models of different sizes and across multiple datasets demonstrate that ATTC effectively reduces the "Tool Ignored" issue, resulting in a performance increase of 4.1% to 7.5%.