When Are Two SAE Latents the Same Feature?
cs.LG
Submitted: 2026-11-03
Updated: 2026-11-03
Terminology
Sources
- Git Re-Basin: Merging Models modulo Permutation Symmetries
- Mechanistic Permutability: Match Features Across Layers
- Semantic Optimal Transport for Sparse Autoencoder Feature Matching and Circuit Compression
- A is for Absorption: Studying Feature Splitting and Absorption in Sparse Autoencoders
- On the Limits of Sparse Autoencoders: A Theoretical Framework and Reweighted Remedy
- Sparse Autoencoders Find Highly Interpretable Features in Language Models
- TinyStories: How Small Can Language Models Be and Still Speak Coherent English?
- Not All Language Model Features Are One-Dimensionally Linear
- Scaling and evaluating sparse autoencoders
- The Platonic Representation Hypothesis
- Transformer Fusion with Optimal Transport
- C$^{2}$R: Cross-sample Consistency Regularization Mitigates Feature Splitting and Absorption in Sparse Autoencoders
- Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders
- Analyze Feature Flow to Enhance Interpretation and Steering in Language Models
- Sparse Autoencoders Do Not Find Canonical Units of Analysis
- Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2
- Towards Principled Evaluations of Sparse Autoencoders for Interpretability and Control
- Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Models
- Model Fusion via Optimal Transport
- Dense SAE Latents Are Features, Not Bugs
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis