提出轻量级注意力一致性评分，高效评估大模型输入词重要性。
Attention Consistency for LLMs Explanation
- 基于最大注意力一致性设计新评分方法
- 相比复杂方法降低22%显存占用，延迟减少30%
- 适合需要高效解释的模型部署场景
理解大语言模型的决策过程对构建可信系统至关重要。然而现有可解释性方法常面临分辨率低、计算开销大的问题。为此，我们提出一种新型、轻量且易于部署的启发式方法——多层注意力一致性评分（MACS），用于估计解码器型模型中输入词的重要程度。MACS通过衡量最大注意力的一致性来评估输入词贡献。实验表明，MACS在可解释质量与计算效率间取得良好平衡，其忠实度接近复杂方法，同时实现22%的显存下降和30%的延迟降低。
原文摘要 · Abstract (English)
Understanding the decision-making processes of large language models (LLMs) is essential for their trustworthy development and deployment. However, current interpretability methods often face challenges such as low resolution and high computational cost. To address these limitations, we propose the \textbf{Multi-Layer Attention Consistency Score (MACS)}, a novel, lightweight, and easily deployable heuristic for estimating the importance of input tokens in decoder-based models. MACS measures contributions of input tokens based on the consistency of maximal attention. Empirical evaluations demonstrate that MACS achieves a favorable trade-off between interpretability quality and computational efficiency, showing faithfulness comparable to complex techniques with a 22\% decrease in VRAM usage and 30\% reduction in latency.
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。