Attention Sinks Induce Gradient Sinks: Massive Activations as Gradient Regulators in Transformers

arXiv:2603.17771 · cs.LG, cs.AI · Submitted 2026-03-18 · Read on arXiv

cs.LG, cs.AI

Submitted: 2026-03-18

Updated: 2026-10-19

Terminology

• Transformer — 3× in this paper · explained in 2 episodes
• Transformers — 2× in this paper · explained in 2 episodes
• Transformer Model — 1× in this paper · explained in 2 episodes
• Transformer Models — 1× in this paper · explained in 2 episodes

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