用大模型迭代解释视觉模型神经元，发现新概念并生成可视化解释。
LINE: LLM-based Iterative Neuron Explanations for Vision Models
- 基于大模型和文生图工具，黑盒迭代优化神经元概念标签。
- 在ImageNet上AUC提升0.11，发现27%预设词表外的新概念。
- 可评估神经元多义性，生成媲美梯度方法的可视化解释。
理解深度神经网络中单个神经元的含义是揭示其复杂决策机制、保障AI安全的关键步骤。尽管神经元标注已有进展，现有方法常受限于预定义的概念词汇表，或产生过于具体的描述，难以捕捉更高阶的全局概念。我们提出LINE，一种无需训练、面向开放词汇表的视觉模型神经元迭代解释方法。在严格黑箱设置下，LINE利用大语言模型与文本到图像生成器，通过激活历史引导，在闭环中迭代提出并优化概念。LINE在多个模型架构上达到领先性能，在ImageNet上AUC提升最高达0.11，在Places365上提升0.05，平均发现27%预设词汇表外的新概念。除识别主要概念外，还提供完整生成历史，支持多义性评估，并生成媲美依赖梯度的激活最大化方法的视觉解释。代码即将开源。
原文摘要 · Abstract (English)
Interpreting individual neurons in deep neural networks is a crucial step towards understanding their complex decision-making processes and ensuring AI safety. Despite recent progress in neuron labeling, existing methods often limit the search space to predefined concept vocabularies or produce overly specific descriptions that fail to capture higher-order, global concepts. We introduce LINE, a novel, training-free iterative approach tailored for open-vocabulary concept labeling in vision models. Operating in a strictly black-box setting, LINE leverages a large language model and a text-to-image generator to iteratively propose and refine concepts in a closed loop, guided by activation history. LINE achieves state-of-the-art performance across multiple model architectures, yielding AUC improvements of up to 0.11 on ImageNet and 0.05 on Places365, while discovering, on average, 27% of new concepts missed by predefined vocabularies. Beyond identifying the top concept, LINE provides a complete generation history, enabling polysemanticity evaluation and producing visual explanations that rival gradient-dependent activation maximization methods. The source code will be made available soon.
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。