Increasing the Analysis Surface of Large Language Models
Stephen Brennan, Ulrich
BSidesSF 2026 · Day 1 · AMC Theatre 14
In an era increasingly defined by the ubiquitous integration of Large Language Models (LLMs), understanding and securing these complex systems has become a paramount challenge. This talk, "Increasing the Analysis Surface of Large Language Models," presented by Stephen Brennan and Ulrich, delves into the inherent difficulties of moderating and securing LLMs due to their statistical nature and focus on syntax over conceptual meaning. It highlights the critical limitations of traditional security approaches, such as input/output filters, and introduces an innovative "white-box" analysis framework known as FORT (Framework for Operational Resilience and Trust).
AI review
Legitimate research direction with real results — white-box attention analysis catching malicious prompts at 78-98% accuracy without I/O filters is a concrete, testable claim worth attention. But the talk as described spends too much runway on transformer 101 and not enough on the methodology that earned those numbers, leaving the most interesting part underexplored.