method · first seen 4 Oct 2026
sparse autoencoder
Also: SAE, crosscoder, sparse feature decomposition, sparse autoencoders for interpretability
Autoencoder with a sparsity penalty used to decompose neural network activations into interpretable features.
Sparse autoencoders decompose a model's internal activations into a larger dictionary of sparsely active directions that are meant to be individually interpretable. In this sample the method is used as a standard tool rather than introduced. One paper replaces per-layer training with model-scope feature discovery. Another builds a structured variant for time-series forecasters. Several others use SAE features for safety-alignment data selection, activation steering, over-refusal mitigation and medical retrieval explanations. A further paper tests how stable feature-recovery claims are. On the code side there is a Google DeepMind repo for mechanistically grounded AI-text detectors and an open SAE research infrastructure project. Nearly all of the activity is on arXiv.
SAEs have been a core mechanistic-interpretability method since about 2023, so the recent mentions reflect steady use rather than a new idea. What is worth tracking is the move toward cross-layer variants, applications in safety and steering, and work testing whether SAE-based claims hold up.
Analyst note (established), 17 Nov 2026.
Mentions per week
PapersCodeDiscussion
Rank 96 on 14 Nov 2026 · 63 mentions in 28 days · 63 independent groups · papers → code · first seen 6 weeks ago · acceleration +0.25
Earliest sources
- ScopeSAE: Model-Scope Feature Discovery with Interpretable Layer SelectionPapers · 4 Oct 2026 · Qingwen Zeng, Zehao Fu, Shuyu Meng, Linghan Huang, Jiayi Zhang, Chenglin Wu, Lin…
- TSAE: Structured Sparse Autoencoders for Interpreting Time-Series Forecasting ModelsPapers · 4 Oct 2026 · Baoxi Liu, Yi Xie
- EmoRSS: Mitigating Emotion-Induced Over-Refusal in Large Language ModelsPapers · 4 Oct 2026 · Shuyi Miao, Yaojin Ma, Chenhang Cui, Xiaohao Liu, Dang Jisheng, Shengda Zhuo, Fe…
- Topology-Conditioned Backdoors: Language Models That Insert Vulnerabilities When They Infer They Are in a Multi-Agent SystemPapers · 5 Oct 2026 · Keegan Wang, Anantika Mannby
- A Testable Theory of Atomic FeaturesPapers · 5 Oct 2026 · Kenny Peng, Jon Kleinberg, Nikhil Garg
- Backdooring Sparse AutoencodersPapers · 5 Oct 2026 · Enrico Ahlers, Daniel Passon, Tobias Kiecker, Eik Reichmann, Lars Grunske
Latest sources
- Beyond Job Titles: Interpreting Early-Career Job Advertisements with Sparse Autoencoders and Semantic ClusteringPapers · 16 Nov 2026 · Yisong Cao, Leah Kim, Brian Wright, Ali Rivera
- Representational Change Under Self-Consuming Training: A Sparse Autoencoder AnalysisPapers · 16 Nov 2026 · Suchit Gupte, Zhongteng Cai, Xueru Zhang, Mohammad Mahdi Khalili
- Interpretable deep concepts of organoid morphologyPapers · 16 Nov 2026 · Lion Jasper Gleiter, Tingying Peng
- Guided Group Featurizers (GGF): Improving Block-Sparse Featurizers with Object-Guided LearningPapers · 13 Nov 2026 · Waldemar Chang, Andrii Krutsylo
- Right Token, Right Time: From Per-Token Gating to Trajectory-Aware Activation SteeringPapers · 13 Nov 2026 · Nima Eshraghi, Lovedeep Gondara, Sagarika Suresh, Xiaotong Xu, Jithin Pradeep, F…
- Explanations that Transfer: Interpreting Medical Retrieval with Sparse AutoEncodersPapers · 13 Nov 2026 · Diane Bouchacourt, Gaëtan Brison
- Deep Latents, Long Memory: Latent persistence explains temporal correlations in transformer activationsPapers · 13 Nov 2026 · Francesco Cagnetta, Santiago Acevedo
- DECODE: Diagnosing Interpretability Failures in Automated Interpretation ProcessPapers · 13 Nov 2026 · Dohyun Kim, Jaeung Lee, Donghee Hong, Jaemin Jo
- When can an interpretability claim be trusted? Stability and certification of feature dimensionality, ridge leverage and the feature recovery ratePapers · 12 Nov 2026 · Ali Tajfar, Behnam Tajfar
- Data-Efficient LLM Safety Alignment Guided by Interpretable FeaturesPapers · 12 Nov 2026 · Hongyu Li, Zhan Yu, Zhe Zhao, Kimmo Yan, Xiong Kuang, Renfen Hu
- Representational Atoms in Neural NetworksPapers · 10 Nov 2026 · Chenhui Hu, Pengfei Cao, Kang Liu, Jun Zhao
- Massive Activations Confound Sparse AutoencodersPapers · 10 Nov 2026 · Nathan Delisle, Chenhao Tan