Publications by Matt Fredrikson
Conference
AGENTHARM: A BENCHMARK FOR MEASURING HARMFULNESS OF LLM AGENTS
Conference
ALIGNED LLMS ARE NOT ALIGNED BROWSER AGENTS
Conference
A RECIPE FOR IMPROVED CERTIFIABLE ROBUSTNESS
Conference
Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization
Conference
Improving Alignment and Robustness with Circuit Breakers
Conference
ON THE PERILS OF CASCADING ROBUST CLASSIFIERS
Conference
CONSISTENT COUNTERFACTUALS FOR DEEP MODELS
Journal Article
Degradation Attacks on Certifiably Robust Neural Networks
Journal Article
Enhancing the insertion of NOP instructions to obfuscate malware via deep reinforcement learning