English

Equivalence of Context and Parameter Updates in Modern Transformer Blocks

Machine Learning 2025-12-24 v2

Abstract

Recent research has established that the impact of context in a vanilla transformer can be represented implicitly by forming a token-dependent, rank-1 patch to its MLP weights. This work extends that foundational theory to the diverse architectures of modern Large Language Models. We first demonstrate a precise, analytical solution for a Gemma-style transformer block, proving that the entire effect of a context can be perfectly mapped to rank-1 patches on its MLP weight matrices and a patch to the RMSNorm scale. We then generalize this result, providing a constructive proof and algorithm for multi-layer models. To unify these findings, we introduce a general framework centered on two core properties: input controllability and output controllability. We prove that a perfect implicit weight patch is possible for any MLP block where the inner function is input-controllable and the outer function is output-controllable. This provides a simpler and more powerful lens for understanding how transformer models transmute prompts into effective weights. This setup generalizes to a wide range of modern LLM architectures including gating, pre-/post-norm, mixture of experts and sequential/parallel transformer blocks.

Keywords

Cite

@article{arxiv.2511.17864,
  title  = {Equivalence of Context and Parameter Updates in Modern Transformer Blocks},
  author = {Adrian Goldwaser and Michael Munn and Javier Gonzalvo and Benoit Dherin},
  journal= {arXiv preprint arXiv:2511.17864},
  year   = {2025}
}