Superposition

Follow this topic Ask about this topic

Recent momentum

-67%

1 papers in the last 28 days · 0.0% of indexed attention

Twelve weeks of publication activity for this topic as it is defined today.

Weekly history

Recent digests

What was published in this topic, kept on the site without email delivery.

Period ending 2026-09-14

1 new paper

A weekly snapshot of new work published in Superposition.

25 papers

Latest in Superposition

Open your feed →

High-probability guarantees for linear accessibility in feature superposition
Sep 10, 2026Enrico VompaSuperposition Sparse Autoencoders
Enhancing Visual Domain Robustness in Behaviour Cloning via Saliency-Guided Augmentation
Aug 12, 2026Zheyu Zhuang, Ruiyu Wang, Nils Ingelhag +2Behavior Cloning Data Augmentation
A Rate Separation for Agnostic Direct Sums
Aug 7, 2026Mihir More, Aritra Das, Debayan GuptaLearning Rates Superposition
Mobius Learning: Cyclic Depth Folding in Transformers
Jul 20, 2026Tongtian ZhuDiffusion Transformers Recursion
From Geometric Recovery to Causal Validation: A Reproducible Audit of Sparse Autoencoder Features, from Superposition Geometry to Causal Inertness
Jul 13, 2026Mohamed Abdessalem BalImproving Sparse Autoencoders Sparse Autoencoders
Compressed Computation under $L^4$ Loss is likely Computation in SuperpositionCompressed Computation under L^4 Loss is likely Computation in Superposition
Jul 6, 2026Francisco Ferreira da Silva, Stefan HeimersheimSuperposition
Resolving superposition in AI for interpretability and cross-modal alignment in patient-neuronal images
Jun 30, 2026Jisung Park, Seohyeon Kang, Daeun Yoo +8Superposition Neuroimaging Models
PairSAE: Mechanistic Interpretability from Pair Representations in Protein Co-Folding
Jun 25, 2026Giosue Migliorini, Aristofanis Rontogiannis, Grigori Guitchounts +3Protein Sparse Autoencoders
Evidence for feature-specific error correction in LLMs
Jun 23, 2026Francisco Ferreira da Silva, Stefan HeimersheimLarge Language Models Fail Error Correction
GEMS: Geometric Constraints Enable Multi-Semantic Superposition in LLMs
Jun 18, 2026Yu DengLinear Activation Steering Steering
Effects of sparsity and superposition on loss in simple autoencoders
Jun 16, 2026Mriganka Basu Roy Chowdhury, Eric McLaughlin WeinerSuperposition Autoencoder
Compressed Computation is (probably) not Computation in Superposition
Jun 12, 2026Jai Bhagat, Sara Molas-Medina, Giorgi Giglemiani +1Superposition Convolutional Neural Network Compression
SuperThoughts: Reasoning Tokens in Superposition
Jun 11, 2026Zheyang Xiong, Shivam Garg, Max Yu +4Chain-Of-Thought Thinking
Post-AGI Economies: Superposition and the Second Fundamental Theorem of Welfare Economics
Jun 6, 2026Elija PerrierWelfare Economies
Vision-Language Asymmetry in Bistable Image Captioning
Jun 6, 2026Arohan AgateDense Video Captioning Flips
Principled Reflection Separation via Nonlinear Superposition and Feature Interaction
Jun 1, 2026Qiming Hu, Mingjia Li, Yuntong Li +1Reflectance Source Separation
Probing for Representation Manifolds in Superposition
May 18, 2026Alexander ModellStatistical Manifold Manifolds
Beyond Linear Superposition: Discovering Climate Features in AI Weather Models with KAN-SAE
May 17, 2026Minjong CheonArtificial Intelligence Weather Models Sparse Autoencoders
Emergence of Frontier Superposition: Möbius attractor and Cascade Supervision
May 12, 2026Hongyu Gu, Jingwen FuSuperposition Downstream Fine-Tuning
Efficient Pre-Training with Token Superposition
May 7, 2026Bowen Peng, Théo Gigant, Jeffrey QuesnelleModel Pretraining Superposition
Superposition Is Not Necessary: A Mechanistic Interpretability Analysis of Transformer Representations for Time Series Forecasting
May 6, 2026Alper YıldırımTime Series Forecasting Sparse Autoencoders
When Attention Collapses: Residual Evidence Modeling for Compositional Inference
May 4, 2026Niklas HoubaModel Attention Superposition
Linear-Readout Floors and Threshold Recovery in Computation in Superposition
May 2, 2026Hector Borobia, Elies Seguí-Mas, Guillermina Tormo-CarbóSuperposition Readout
Structural Instability of Feature Composition
Apr 18, 2026Yunpeng ZhouSparse Autoencoders Superposition
Similarity of Neural Network Representations in Superposition
Mar 31, 2026Sunny Liu, Habon Issa, André Longon +4Neural Representations Superposition

Sep 10, 2026 · stat.MLJ/K move · Enter open · S save

High-probability guarantees for linear accessibility in feature superposition

Enrico Vompa

Neural networks can leverage feature superposition to encode more concepts than dimensions, but cross-feature interference constrains the linear accessibility of simultaneously active features. By framing linear accessibility as a compressed sensing problem, we derive high-probability bounds for fixed supports under subgaussian noise, proving the sufficient dimension scales linearly (

d=O_{\varepsilon}(k \log m)

) rather than prior worst-case quadratic limits. We then validate these bounds across system parameters through Gaussian-tail approximations. These results quantify the geometric constraints of the linear representation hypothesis, providing a framework for evaluating sparse autoencoders, compositional generalization, and neural interpretability.

Superposition Sparse Autoencoders Concentration Inequalities Interference

Open paper Ask about it PDF