Efficient Reasoning with Balanced Thinking

Abstract

Large Reasoning Models (LRMs) have shown remarkable reasoning capabilities, yet they often suffer from overthinking, expending redundant computational steps on simple problems, or underthinking, failing to explore sufficient reasoning paths despite inherent capabilities. These issues lead to inefficiencies and potential inaccuracies, limiting practical deployment in resource-constrained settings. Existing methods to mitigate overthinking, such as suppressing reflective keywords or adjusting reasoning length, may inadvertently induce underthinking, compromising accuracy. Therefore, we propose ReBalance, a training-free framework that achieves efficient reasoning with balanced thinking. ReBalance leverages confidence as a continuous indicator of reasoning dynamics, identifying overthinking through high confidence variance and underthinking via consistent overconfidence. By aggregating hidden states from a small-scale dataset into reasoning mode prototypes, we compute a steering vector to guide LRMs' reasoning trajectories. A dynamic control function modulates this vector's strength and direction based on real-time confidence, pruning redundancy during overthinking, and promoting exploration during underthinking. Extensive experiments conducted on four models ranging from 0.5B to 32B, and across nine benchmarks in math reasoning, general question answering, and coding tasks demonstrate that ReBalance effectively reduces output redundancy while improving accuracy, offering a general, training-free, and plug-and-play strategy for efficient and robust LRM deployment. Code is available at https://github.com/yu-lin-li/ReBalance .

Figures (16)

• Figure 1: Qualitative and quantitative comparisons with previous state-of-the-art methods for mitigating overthinking. (a) Given the question "For what real values of $x$ is $-4 < x^{4} + 4x^{2} < 21$?", the model first obtains intervals $(-\sqrt{3}, 0)$ and $(0, \sqrt{3})$, and then verifies if $x = 0$ is included. However, the baseline deepseek_r1 redundantly checks irrelevant values after correctly validating $x = 0$, causing overthinking. Current mitigation methods deer overly suppress necessary reflection, leading to underthinking. Our method dynamically controls the reasoning state, effectively balancing these two extremes. (b)ReBalance outperforms previous state-of-the-art method seal across multiple mathematical reasoning datasets and model scales (0.5B–32B), reducing reasoning length while simultaneously improving accuracy.
• Figure 2: (a) Effects of overthinking mitigation on reasoning modes. We compare the distributions of reasoning lengths for correct and incorrect predictions before and after applying overthinking mitigation methods. The reduction in reasoning lengths for correct and incorrect predictions indicates the degree to which overthinking is alleviated and underthinking is introduced, respectively. Existing methods significantly introduce underthinking, whereas our method effectively achieves a balanced reduction of both. (b) Correlation between confidence and reasoning modes. We observe that the overthinking samples exhibit higher confidence variance compared to normal samples, while underthinking samples show persistently high confidence levels.
• Figure 3: Illustration of the ReBalance framework. We first perform offline one-pass data collection on a small-scale seen dataset. At each step, the steering vector is extracted at the first token of the specified layer based on confidence, and a dynamic function is fitted according to model behaviors. During deployment, the dynamic function outputs steering weights based on the model's real-time confidence online, thus balancing between overthinking and underthinking.
• Figure 4: (a-b) Layerwise Performance of MATH-500 for (a) R1–7B and (b) QwQ–32B.(c) Sensitivity to sample size for steering vector extraction.(d) Performance with cross-domain vectors.
• Figure 5: Static $\alpha_a$ Control on MATH-500. Top: R1-7B; Bottom: QwQ--32B.