Applications
The Role of Feed-Forward Layers in Transformer Dynamics
arXiv:2609.36230v1 Announce Type: new Abstract: We study the dynamical behavior of tokens in transformers from a control-theoretic perspective. Our model includes the feed-forward layer present after
arXiv:2609.36230v1 Announce Type: new Abstract: We study the dynamical behavior of tokens in transformers from a control-theoretic perspective. Our model includes the feed-forward layer present after the self-attention mechanism, with the self-attention mechanism interpreted as an interacting particle system and the feed-forward layer as an independent control. Our main theoretical result establishes that the feed-forward network can steer the tokens arbitrarily close to consensus regardless of the key, query, and value matrices. Our result are easily extended to convergence to many clusters and to multi-head attention. We conduct numerical experiments to verify our results, and compare thresholding behavior from our theory to real-world LLMs.
Related
- Gradient Flow Structure and Quantitative Dynamics of Multi-Head Self-Attention
- Unified Simulation of Lagrangian Particle Dynamics via Transformer
- Temporal Contrastive Transformer for Financial Crime Detection: Self-Supervised Sequence Embeddings via Predictive Contrastive Coding
Source: arXiv cs.LG | 2026-09-30