This inventor holds 1 USPTO granted patent. Top assignee: Cato Networks Ltd. Active years: 2026.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2026
Title: Ofir Abu and Network Security for Machine Learning Models
Introduction
Ofir Abu is an inventor based in Tel Aviv, Israel, and works at Cato Networks Ltd. His patent work addresses security risks in AI applications guided by large language models (LLMs).
Latest Patents
Abu is associated with 1 patent, titled “Network Security for Machine Learning Models Including Large Language Models (LLMs).” The patent describes a system that traces control or data flow through a model’s transformer attention heads during an interaction. It identifies output tokens that stem from input tokens supplied by external resources and can trigger a security action to prevent those resources from feeding adversarial input into the AI application.
Career Highlights
Abu’s listed work at Cato Networks Ltd. connects his invention to the challenge of protecting LLM-guided applications from adversarial inputs originating outside the application.
Collaborations
His listed coworkers include Itay Ravia and Adir Gruss.
Conclusion
Ofir Abu’s patent focuses on tracing how external inputs influence an LLM’s outputs and using that insight to help prevent adversarial input from compromising AI applications.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile