Tomer Wetzler
Security researcher & engineer - started my security journey doing research on open source software security, and developing the tools to scan and defend against them. Then, pivoted to cloud apps and identity research in a full research scope, where I researched and created rules to mitigate complex attacks and the use of various attack tools. Now I'm focusing on AI security, where I work to defend AI agents everywhere!
Posts by Tomer Wetzler
Catching Prompt Guard Off Guard: Exploiting Overfit in Training Algorithms
How understanding the training algorithms used in machine learning models may allow attacker to bypass them entirely
Tomer Wetzler
Moving The Decision Boundary of LLM Safety Classifiers
How a new fine-tuning approach can mitigate the problem of inaccurate safety paths
Tomer Wetzler
The Geometry of Safety Failures in Large Language Models
A deep dive into activation space of prompts in safety classifiers. Showing not why - but where - safety fails in LLM classifiers meant to detect malicious prompts.
Tomer Wetzler
Enabling Safety in AI Agents via Choice Architecture
How adding a single safety labeled tool to an LLM's toolset can sharply increase its defense
Tomer Wetzler
Modeling LLMs via Structured Self-Modeling (SSM)
How using structured prompts present findings of self-modeling in LLMs, which may benefit both attackers and defenders
Tomer Wetzler
Data-Structure Injection (DSI) in AI Agents
How controlling the structure of the prompt, not just the semantics, can exploit your AI agents and their tools
Tomer Wetzler