Skip to main content
Home/Risks/Cui et al. (2024)/Memory and Storage

Memory and Storage

Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems

Cui et al. (2024)

Sub-category
Risk Domain

Vulnerabilities that can be exploited in AI systems, software development toolchains, and hardware, resulting in unauthorized access, data and privacy breaches, or system manipulation causing unsafe outputs or behavior.

"Similar to conventional programs, hardware infrastructures can also introduce threats to LLMs. Memory-related vulnerabilities, such as rowhammer attacks [160], can be leveraged to manipulate the parameters of LLMs, giving rise to attacks such as the Deephammer attack [167], [168]."(p. 10)

Part of Hardware Vulnerabilities

Other risks from Cui et al. (2024) (49)