Spectral Outliers Reveal Dominant Learned Structure in Transformer Attention
Kasun Dewage, Marianna Pensky, Suranadi De Silva, T. H. Bandara
cs.LG, cs.AI, cs.CL
Submitted: 2026-08-08
Updated: 2026-08-11
Comments: Accepted at the International Conference on Machine Learning and Applications (ICMLA 2026); to appear in IEEE proceedings
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- In-context Learning and Induction Heads
- A Spectral Condition for Feature Learning
- RoBERTa: A Robustly Optimized BERT Pretraining Approach
- OPT: Open Pre-trained Transformer Language Models
- LLaMA: Open and Efficient Foundation Language Models
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- The Llama 3 Herd of Models
- Mistral 7B
- Qwen2.5 Technical Report
- Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone
- Toy Models of Superposition
- LORA-CRAFT: Cross-layer Rank Adaptation via Frozen Tucker Decomposition of Pre-trained Attention Weights
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