让大模型自己判断该想多深，大幅减少冗余计算。
Aware First, Think Less: Dynamic Boundary Self-Awareness Drives Extreme Reasoning Efficiency in Large Language Models
- 模型动态感知问题难度，自动调节推理深度。
- 响应令牌减少49.27%，训练时间缩短5倍。
- 适合资源受限场景，推理效率远超传统方法。
近期大语言模型在复杂推理任务上通过长链式思维（Long Chain-of-Thought, CoT）取得显著进步，但常伴随大量冗余，影响计算效率并导致实时应用延迟。现有方法依赖人工定义的难度先验，与模型自知难度不匹配，造成效率低下。本文提出动态推理边界自感知框架（DR.SAF），使模型能根据问题复杂度动态评估并调整推理深度。该框架包含边界自感知对齐、自适应奖励管理与边界保持机制，有效优化推理过程，在不牺牲性能的前提下实现高效平衡。实验表明，DR.SAF将总响应令牌减少49.27%，提升6.59倍令牌效率，训练时间减少5倍，适用于资源受限环境。极端训练下，其令牌效率超越传统指令模型，准确率还高出16%以上。
原文摘要 · Abstract (English)
Recent advancements in large language models (LLMs) have greatly improved their capabilities on complex reasoning tasks through Long Chain-of-Thought (CoT). However, this approach often results in substantial redundancy, impairing computational efficiency and causing significant delays in real-time applications. To improve the efficiency, current methods often rely on human-defined difficulty priors, which do not align with the LLM's self-awared difficulty, leading to inefficiencies. In this paper, we introduce the Dynamic Reasoning-Boundary Self-Awareness Framework (DR. SAF), which enables models to dynamically assess and adjust their reasoning depth in response to problem complexity. DR. SAF integrates three key components: Boundary Self-Awareness Alignment, Adaptive Reward Management, and a Boundary Preservation Mechanism. These components allow models to optimize their reasoning processes, balancing efficiency and accuracy without compromising performance. Our experimental results demonstrate that DR. SAF achieves a 49.27% reduction in total response tokens with minimal loss in accuracy. The framework also delivers a 6.59x gain in token efficiency and a 5x reduction in training time, making it well-suited to resource-limited settings. During extreme training, DR. SAF can even surpass traditional instruction-based models in token efficiency with more than 16% accuracy improvement.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。