The paper investigates Sparse Autoencoders (SAEs) by applying dictionary learning identifiability results to derive constraints on optimal dictionary learning features. It shows that optimal features cannot activate simultaneously, explaining SAE oddities such as hierarchical splitting, absorption, dense antipodal features, and infinite feature splitting as consequences of the dictionary learning objective. The derived constraints also serve as diagnostics, revealing that real SAEs perform well on training data but fail on increasingly distant test data.
By William Dorrell
arXiv:2606. 06333v1 Announce Type: new Abstract: Sparse Autoencoders (SAEs) are widely used for mechanistic interpretability in large language models, yet their formulation assigns each latent feature a single decoder direction, implicitly assuming features to be one-dimensional.
By Seyed Arshan Dalili, Mehrdad Mahdavi
arXiv:2606. 11722v1 Announce Type: cross Abstract: Finding interpretable directions in language-model representations is critical for understanding and controlling model behavior.
By Sida Liu, Feijiang Han
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.
By Haoran Jin, Xiting Wang, Shijie Ren, Hong Xie, Defu Lian
arXiv:2606. 18383v1 Announce Type: new Abstract: Sparse autoencoders (SAEs) are increasingly used to extract interpretable features from language models (LMs), yet a central question remains: when can an SAE-based explanation be treated as a faithful view of an underlying frozen LM We study this through a post-hoc generalization framework that certifies the LM via a sparse proxy, obtained by replacing a native hidden activation with its pretrained SAE reconstruction.
By Dibyanayan Bandyopadhyay, Asif Ekbal
arXiv:2607. 20652v1 Announce Type: cross Abstract: Language models are thought to exhibit the phenomenon of superposition, representing many more features than dimensions in their residual streams.
By Andrew Mack, Kraig Yuheng Tou, Mark Henry, Zhengxun Wu, Lauren Greenspan
arXiv:2607. 01799v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) decompose internal activations of neural networks into sparse linear combinations of learned features by fitting an overcomplete dictionary $\mathbf{W}\in\mathbb{R}^{m\times n}$ with $m<n$, and inferring a sparse code $\mathbf{x}\in\mathbb{R}^n$ from $\mathbf{h}\approx\mathbf{W}\mathbf{x}$.
By Rodrigo Mendoza-Smith
arXiv:2606. 16028v1 Announce Type: new Abstract: Modern deep learning architectures are increasingly multi-task and multi-modal, using a pretrained foundation model combined with task-specific, fine-tuned models.
By Thomas Dittrich, Oliver Potocki, Philipp Grohs
arXiv:2508. 16560v4 Announce Type: replace-cross Abstract: Sparse Autoencoders (SAEs) extract features from LLM internal activations, meant to correspond to interpretable concepts.
By David Chanin, Adri\`a Garriga-Alonso
arXiv:2607. 08605v1 Announce Type: cross Abstract: 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.
By Weiduo Liao, Yunqiao Yang, Ying Wei
arXiv:2606. 26620v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) have emerged as a powerful tool for decomposing superposed language model representations into sparse and interpretable features.
By XinYang He, Wei Wang, Bing Zhao, Xuan Ren, WenBo Li, WeiXu Qiao, Hu Wei, Lin Qu
Monotonicity has been a long-running architectural inductive bias for neural networks, motivated by tabular, scientific, and economic settings where outputs are known to respond monotonically to certain inputs. Existing approaches are MLP- or flow-based and lack per-edge functional transparency; the only Kolmogorov--Arnold Network (KAN) variant with monotonicity, MonoKAN, enforces the constraint only on a restricted parameter subset and requires a projection-style training procedure.