Understanding the Security Boundary of Obfuscation-based On-Device LLM Protection
Researchers formalize obfuscation primitives for TEE-protected on-device LLMs and show a Collapse attack breaks ArrowCloak, TSQP, and LoRO, then extend the boundary.
The paper formalizes obfuscation primitives for TEE-Shielded LLM Partition (TSLP) schemes that offload computationally intensive layers from a Trusted Execution Environment to external GPUs. A novel primitive-guided attack, Collapse, demonstrates a shared vulnerability in prominent published methods including ArrowCloak (Security'25), TSQP (S&P'25), and LoRO (NeurIPS'25). The authors then introduce two new obfuscation primitives and integrate them with existing constructs to formulate an extended security boundary (O_ext).
- Formalizes obfuscation primitives unifying prior TEE-Shielded LLM Partition defenses
- Collapse attack breaks ArrowCloak, TSQP, and LoRO via a shared structural vulnerability
- Two new primitives extend the security boundary beyond prior primitive compositions
- Targets protection of on-device LLM intellectual property against specialized adversarial attacks
- Products
- ArrowCloakTSQPLoRO
From the original article
1 min read at arxiv.orgTrusted Execution Environments (TEEs) offer a promising mechanism for safeguarding the intellectual property of on-device Large Language Models (LLMs).
By Hanyi Zhou · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.