Interference Beyond Geometry in Concept Extraction
cs.LG
Submitted: 2026-09-28
Updated: 2026-09-28
Code: https://github.com/decoderesearch/SAELens
Sources
- On the Complexity of Neural Computation in Superposition
- Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling
- Near-Floor Geometry Is Generic: Leverage Dispersion in Trained Overcomplete Codes
- Alternating minimization for dictionary learning: Local Convergence Guarantees
- Effects of sparsity and superposition on loss in simple autoencoders
- How Optimality Structures Sparse Dictionaries: Theory for Interpreting SAE Representations
- Towards A Rigorous Science of Interpretable Machine Learning
- Toy Models of Superposition
- Measuring Monosemanticity in Sparse Autoencoders via Latent Activation Coherence
- Mathematical Models of Computation in Superposition
- Spectral Superposition: A Theory of Feature Geometry
- SAEBench: A Comprehensive Benchmark for Sparse Autoencoders in Language Model Interpretability
- Adam: A Method for Stochastic Optimization
- OrtSAE: Orthogonal Sparse Autoencoders Uncover Atomic Features
- What Causes Polysemanticity? An Alternative Origin Story of Mixed Selectivity from Incidental Causes
- k-Sparse Autoencoders
- Superposition Without Interference? Towards Isolated Interventions via Almost Orthogonal Features in Language Models
- Improving Dictionary Learning with Gated Sparse Autoencoders
- Jumping Ahead: Improving Reconstruction Fidelity with JumpReLU Sparse Autoencoders
- Polysemanticity and Capacity in Neural Networks
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks