BadEngram: Backdoor Attack on Gated Memory Components in LLMs

To expand open-weight models' capacity without proportionally increasing computation, recent language models incorporate gated parametric memories that retrieve learned values and inject them into intermediate representations. Despite these efficiency benefits, such modules create a distinct atta...

CVE-2025-37173

An improper input handling vulnerability exists in the web-based management interface of mobility conductors running either AOS-10 or AOS-8 operating systems. Successful exploitation could allow an authenticated malicious actor with valid credentials to trigger unintended behavior on the affected...