When can an interpretability claim be trusted? Stability and certification of feature dimensionality, ridge leverage and the feature recovery rate
cs.LG, cs.AI
Submitted: 2026-11-12
Updated: 2026-11-12
Terminology
Sources
- Leverage-Weighted Conformal Prediction
- Spectral Superposition: A Theory of Feature Geometry
- Unstable Features, Reproducible Subspaces: Understanding Seed Dependence in Sparse Autoencoders
- Taming Polysemanticity in LLMs: Provable Feature Recovery via Sparse Autoencoders
- Input Sparsity Time Low-Rank Approximation via Ridge Leverage Score Sampling
- Toy Models of Superposition
- Dictionary learning -- from local towards global and adaptive
- Sparse Autoencoders Trained on the Same Data Learn Different Features
- Polysemanticity and Capacity in Neural Networks
- Lipschitz-Margin Training: Scalable Certification of Perturbation Invariance for Deep Neural Networks
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