EntroCut: Entropy-Guided Adaptive Truncation for Efficient Chain-of-Thought Reasoning in Small-scale Large Reasoning Models
This addresses the problem of high computational costs for users of reasoning models, offering an incremental improvement in efficiency.
The paper tackles the computational inefficiency of Large Reasoning Models by proposing EntroCut, an entropy-guided method that dynamically truncates reasoning steps, reducing token usage by up to 40% with minimal accuracy loss.
Large Reasoning Models (LRMs) excel at complex reasoning tasks through extended chain-of-thought generation, but their reliance on lengthy intermediate steps incurs substantial computational cost. We find that the entropy of the model's output distribution in early reasoning steps reliably distinguishes correct from incorrect reasoning. Motivated by this observation, we propose EntroCut, a training-free method that dynamically truncates reasoning by identifying high-confidence states where reasoning can be safely terminated. To comprehensively evaluate the trade-off between efficiency and accuracy, we introduce the Efficiency-Performance Ratio (EPR), a unified metric that quantifies relative token savings per unit accuracy loss. Experiments on four benchmarks show that EntroCut reduces token usage by up to 40\% with minimal accuracy sacrifice, achieving superior efficiency-performance trade-offs compared with existing training-free methods. These results demonstrate that entropy-guided dynamic truncation provides a practical approach to mitigate the inefficiency of LRMs.