通过分析模型中的捷径神经元，实现更可信的LLM评估。
Establishing Trustworthy LLM Evaluation via Shortcut Neuron Analysis
- 通过对比与因果分析识别导致过拟合的捷径神经元。
- 提出捷径神经元修复方法，使评估结果与MixEval相关性超0.95。
- 无需构建新基准，适用于多种模型和参数设置。
大型语言模型的发展依赖于可信的评估。然而，当前多数评估基于公开基准，易受数据污染影响，严重损害公平性。已有研究尝试构建动态基准应对污染，但持续创建新基准成本高且循环频繁。本文旨在通过分析受污染模型的内在机制来解决该问题。实验发现，模型性能被高估的原因在于训练过程中参数习得捷径解。我们提出一种通过比较与因果分析识别捷径神经元的新方法，并基于此设计了捷径神经元修补评估法，以抑制这些神经元。实验验证了该方法在缓解污染方面的有效性。此外，评估结果与近期发布的可信基准MixEval呈现强线性相关，斯皮尔曼系数（ρ）超过0.95，表明该方法能准确揭示模型真实能力，具备可信性。进一步实验验证了方法在不同基准和超参数设置下的泛化能力。
原文摘要 · Abstract (English)
The development of large language models (LLMs) depends on trustworthy evaluation. However, most current evaluations rely on public benchmarks, which are prone to data contamination issues that significantly compromise fairness. Previous researches have focused on constructing dynamic benchmarks to address contamination. However, continuously building new benchmarks is costly and cyclical. In this work, we aim to tackle contamination by analyzing the mechanisms of contaminated models themselves. Through our experiments, we discover that the overestimation of contaminated models is likely due to parameters acquiring shortcut solutions in training. We further propose a novel method for identifying shortcut neurons through comparative and causal analysis. Building on this, we introduce an evaluation method called shortcut neuron patching to suppress shortcut neurons. Experiments validate the effectiveness of our approach in mitigating contamination. Additionally, our evaluation results exhibit a strong linear correlation with MixEval, a recently released trustworthy benchmark, achieving a Spearman coefficient ($ρ$) exceeding 0.95. This high correlation indicates that our method closely reveals true capabilities of the models and is trustworthy. We conduct further experiments to demonstrate the generalizability of our method across various benchmarks and hyperparameter settings. Code: https://github.com/GaryStack/Trustworthy-Evaluation
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。