How Vulnerable Are Edge LLMs?
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866914420389576704 |
|---|---|
| author | Ding, Ao Li, Hongzong Liang, Zi Shi, Zhanpeng Zhuang, Shuxin Tang, Shiqin Feng, Rong Lu, Ping |
| author_facet | Ding, Ao Li, Hongzong Liang, Zi Shi, Zhanpeng Zhuang, Shuxin Tang, Shiqin Feng, Rong Lu, Ping |
| contents | Large language models (LLMs) are increasingly deployed on edge devices under strict computation and quantization constraints, yet their security implications remain unclear. We study query-based knowledge extraction from quantized edge-deployed LLMs under realistic query budgets and show that, although quantization introduces noise, it does not remove the underlying semantic knowledge, allowing substantial behavioral recovery through carefully designed queries. To systematically analyze this risk, we propose \textbf{CLIQ} (\textbf{Cl}ustered \textbf{I}nstruction \textbf{Q}uerying), a structured query construction framework that improves semantic coverage while reducing redundancy. Experiments on quantized Qwen models (INT8/INT4) demonstrate that CLIQ consistently outperforms original queries across BERTScore, BLEU, and ROUGE, enabling more efficient extraction under limited budgets. These results indicate that quantization alone does not provide effective protection against query-based extraction, highlighting a previously underexplored security risk in edge-deployed LLMs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_23822 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | How Vulnerable Are Edge LLMs? Ding, Ao Li, Hongzong Liang, Zi Shi, Zhanpeng Zhuang, Shuxin Tang, Shiqin Feng, Rong Lu, Ping Cryptography and Security Computation and Language Machine Learning Large language models (LLMs) are increasingly deployed on edge devices under strict computation and quantization constraints, yet their security implications remain unclear. We study query-based knowledge extraction from quantized edge-deployed LLMs under realistic query budgets and show that, although quantization introduces noise, it does not remove the underlying semantic knowledge, allowing substantial behavioral recovery through carefully designed queries. To systematically analyze this risk, we propose \textbf{CLIQ} (\textbf{Cl}ustered \textbf{I}nstruction \textbf{Q}uerying), a structured query construction framework that improves semantic coverage while reducing redundancy. Experiments on quantized Qwen models (INT8/INT4) demonstrate that CLIQ consistently outperforms original queries across BERTScore, BLEU, and ROUGE, enabling more efficient extraction under limited budgets. These results indicate that quantization alone does not provide effective protection against query-based extraction, highlighting a previously underexplored security risk in edge-deployed LLMs. |
| title | How Vulnerable Are Edge LLMs? |
| topic | Cryptography and Security Computation and Language Machine Learning |
| url | https://arxiv.org/abs/2603.23822 |