Abstract
Activation steering manipulates large language model behavior by intervening on internal activations, but the mechanistic basis of these interventions remains poorly understood. We decompose refusal steering into component-level interventions across four open-weight models, identifying the sparse subsets of attention and MLP components whose steering suffices to reproduce the full behavioral effect. We find that refusal directions concentrate in sparse component mechanisms comprising 28--48\% of upstream components, retaining 88--101\% of steering effectiveness. Within these mechanisms, effective steering further concentrates in approximately 50\% of residual stream dimensions, retaining 85--98\% of the component-mechanism baseline, consistent with a privileged basis structure. Sparsity thus operates at two levels: which components are steered, and which dimensions within those components carry the signal. Together these findings show that refusal is not diffusely encoded across a transformer but assembled by a structured, identifiable mechanism, providing a foundation for mechanistic understanding of how refusal behaviors are represented and steered. To facilitate reproducibility, we release all code and raw experimental results in https://github.com/wang-research-lab/Refusal_Mechanisms.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Siu, V., Grant-Richards, G., Pavlovich, V., Sun, Y., Song, D., & Wang, C. (2026). Component and Dimension Sparsity in Transformer Refusal Mechanisms. https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms
MLA 9
Siu, Vincent, et al. "Component and Dimension Sparsity in Transformer Refusal Mechanisms." https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms.
Chicago (author–date)
Siu, Vincent, Glenn Grant-Richards, Vlad Pavlovich, Yizhou Sun, Dawn Song, and Chenguang Wang. 2026. "Component and Dimension Sparsity in Transformer Refusal Mechanisms." https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms.
Harvard
Siu, V., Grant-Richards, G., Pavlovich, V., Sun, Y., Song, D. and Wang, C. (2026) 'Component and Dimension Sparsity in Transformer Refusal Mechanisms', Available at: https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms.
Vancouver
Siu V, Grant-Richards G, Pavlovich V, Sun Y, Song D, Wang C. Component and Dimension Sparsity in Transformer Refusal Mechanisms. https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms
IEEE
V. Siu, G. Grant-Richards, V. Pavlovich, Y. Sun, D. Song, and C. Wang, "Component and Dimension Sparsity in Transformer Refusal Mechanisms," https://omanscience.com/en/articles/component-and-dimension-sparsity-in-transformer-refusal-mechanisms.