Efficient Reasoning for Large Reasoning Language Models via Certainty-Guided Reflection Suppression

TL;DR

This paper introduces Certainty-Guided Reflection Suppression (CGRS), a method to reduce redundant reasoning steps in large reasoning language models, significantly decreasing token usage while maintaining reasoning accuracy.

Contribution

CGRS is a model-agnostic, no-retraining approach that dynamically suppresses reflection triggers based on confidence, improving efficiency in large reasoning language models.

Findings

Reduces token usage by 18.5% to 41.9%.

Maintains reasoning accuracy across benchmarks.

Effective across various model architectures and scales.

Abstract

Recent Large Reasoning Language Models (LRLMs) employ long chain-of-thought reasoning with complex reflection behaviors, typically signaled by specific trigger words (e.g., "Wait" and "Alternatively") to enhance performance. However, these reflection behaviors can lead to the overthinking problem where the generation of redundant reasoning steps that unnecessarily increase token usage, raise inference costs, and reduce practical utility. In this paper, we propose Certainty-Guided Reflection Suppression (CGRS), a novel method that mitigates overthinking in LRLMs while maintaining reasoning accuracy. CGRS operates by dynamically suppressing the model's generation of reflection triggers when it exhibits high confidence in its current response, thereby preventing redundant reflection cycles without compromising output quality. Our approach is model-agnostic, requires no retraining or architectural modifications, and can be integrated seamlessly with existing autoregressive generation pipelines. Extensive experiments across four reasoning benchmarks (i.e., AIME24, AMC23, MATH500, and GPQA-D) demonstrate CGRS's effectiveness: it reduces token usage by an average of 18.5% to 41.9% while preserving accuracy. It also achieves the optimal balance between length reduction and performance compared to state-of-the-art baselines. These results hold consistently across model architectures (e.g., DeepSeek-R1-Distill series, QwQ-32B, and Qwen3 family) and scales (4B to 32B parameters), highlighting CGRS's practical value for efficient reasoning.