Abstract
As small language models (SLMs) are increasingly deployed on resource-constrained and on-device platforms, including as components of agentic systems, the integrity of locally stored model parameters becomes an important safety concern. We investigate whether safety-sensitive behavior in LLaMA-2-7B-Chat is concentrated within a sparse subset of parameters, creating a reduced fault surface for targeted analysis. We study two complementary localization methods: low-rank safety-associated subspace analysis and parameter-level safety--utility importance filtering. Both approaches reveal highly non-uniform safety sensitivity across the network, with the MLP down_proj consistently emerging as a prominent safety-sensitive component and o_proj providing a smaller contribution. Using parameter-level localization, modifying only 0.19% of model weights in down_proj yields 53% Basic ASR and 56% GCG ASR, while tinyBenchmarks accuracy remains at 51.6% compared with a 52.2% unmodified baseline. These results motivate targeted fault analysis and selective integrity protection for language models deployed in resource-constrained, on-device, and agentic settings.
Keywords
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Karamat, M. Z., & Garcia, C. C. (2026). How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis. https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis
MLA 9
Karamat, Muhammad Zeeshan, and Christiana Chamon Garcia. "How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis." https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis.
Chicago (author–date)
Karamat, Muhammad Zeeshan, and Christiana Chamon Garcia. 2026. "How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis." https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis.
Harvard
Karamat, M. Z. and Garcia, C. C. (2026) 'How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis', Available at: https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis.
Vancouver
Karamat MZ, Garcia CC. How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis. https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis
IEEE
M. Z. Karamat, and C. C. Garcia, "How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis," https://omanscience.com/en/articles/how-fragile-is-on-device-language-model-safety-localizing-safety-critical-parameters-for-sparse-fault-analysis.