这篇论文提的S^2AE让视觉语言模型里稀疏编码的概念更一致，比普通SAE语义对齐涨了6%，还能保持高重建质量。搞多模态可解释性的可以看看。
稀疏自编码器(SAE)在机械可解释性中学习稀疏潜在特征，但在视觉-语言模型(VLM)中，香草SAE难以学到跨模态一致的概念。研究者提出结构化稀疏自编码器(S^2AE)，在Qwen2.5-VL-7B-Instruct上评估，语义对齐(mIoU)平均提升6.06%，表征效率(l0范数)降低60.81，重构保真度(解释方差)保持在99%以上。跨模态分析显示S^2AE通过视觉结构先验增强神经元单语义性，语义一致性平均提升3.08%，单语义性得分平均提升2.37%。
When Structured Sparse Autoencoders Learn Consistent Concepts Across Modalities
Sparse autoencoders (SAEs) have emerged as a promising technique for mechanistic interpretability by learning a set of sparse latent features in large models, each of which encodes a distinct concept. However, in vision-language models (VLMs), vanilla SAEs struggle to learn modality-consistent concepts, with concepts often exhibiting fragmented coverage (i.e., disjoint regions) in the visual modality. To address this challenge, we propose a Structured Sparse AutoEncoder ($S^2AE$) that enforces concept consistency from both semantic and spatial perspectives in the visual modality. Specifically, we group image patches based on Transformer attention similarity and spatial proximity, and introduce a structured sparsity regularization when training the vanilla SAE. The regularization consists of exclusive sparsity for inter-group concept disentanglement and group sparsity for intra-group concept consistency, which drives the latent neurons by SAEs to specialize in distinct, semantically grounded concepts. Evaluated on the \texttt{Qwen2.5-VL-7B-Instruct} model, the method achieves 6.06% average improvement in semantic alignment (mIoU) and 60.81 in representational efficiency (lower l0 norm) while maintaining near-perfect reconstruction fidelity with an Explained Variance above 99%. Cross-modal analysis further demonstrates that $S^2AE$ enhances neuronal monosemanticity by this visual structural prior, achieving a 3.08% average gain in semantic consistency and a 2.37% average gain in monosemanticity scores for both modalities of multimodal features, thereby fostering more coherent and disentangled representations.