The Identifiability and Observability of Deep Normalized Attention
cs.LG
Submitted: 2026-10-07
Updated: 2026-10-07
Sources
- Algebraic Invariants of Lightning Self-Attention
- On Identifiability in Transformers
- Functional dimension of feedforward ReLU neural networks
- On the Expressive Power of Deep Polynomial Neural Networks
- Provably Learning Multi-Head Attention with Queries
- Everything, Everywhere, All at Once: Is Mechanistic Interpretability Identifiable?
- Geometry of fibers of the multiplication map of deep linear neural networks
- Attention Is All You Need
- Affine symmetries and neural network identifiability
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