两篇新研究论文解决了使用稀疏自编码器（SAE）解释大型语言模型时面临的挑战。第一篇论文介绍了C$^2$R（跨样本一致性正则化），以缓解特征分裂和吸收问题，这些问题源于跨样本的不一致的潜在分配。第二篇论文识别并解决了视觉-语言模型中的跨模态特征异质性问题，在这种情况下，相同概念根据其在图像或文本嵌入中的表示，可能会激活不同的潜在方向。 AI
影响 这些论文提供了改进AI模型可解释性和可靠性的新技术，有望更好地理解和控制其内部工作机制。
排序理由 两篇在arXiv上发表的学术论文，介绍了用于解释AI模型的新方法。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.30609v1 Announce Type: cross Abstract: Sparse Autoencoders (SAEs) are widely used to interpret large language models by decomposing activations into sparse, human-understandable features, but scaling to large dictionaries exposes fundamental challenges. Systematic stud…
arXiv:2606.29888v1 Announce Type: new Abstract: Vision-language models map images and text into a joint embedding space. However, these embeddings often entangle multiple semantic features, which limits their interpretability and controllability. While sparse autoencoders have em…
Sparse Autoencoders (SAEs) are widely used to interpret large language models by decomposing activations into sparse, human-understandable features, but scaling to large dictionaries exposes fundamental challenges. Systematic studies reveal pervasive feature splitting that fragme…