Research Area
LLM Safety and Evaluation
Measurement of what survives adversarial pressure in LLM-agent behaviour — persistence, structural hardening, masking effects and held-out evaluation designed to resist Goodhart-style overfitting.
Works in this area (2)
-
PreprintEmpirical StudySynthetic Controlled Pilot
Measuring What Survives: Pressure-Filtered Persistence and a Masking Index for LLM-Agent Safety Evaluation
When a safety intervention lowers an agent’s attack-success rate, two different things may have happened: the intervention removed a load-bearing failure route, or the agent learned the measured attacks while the route survived. Frequency metrics are necessary, but they do not by themselves…
-
PreprintEmpirical StudySynthetic Controlled Pilot
A Topological Robustness and Anti-Fragility Framework for LLM Agents
Adversarial evaluation of large-language-model agents is usually expressed as event counts: attack-success rate, policy-violation rate, refusal-failure rate, pass percentage, or robustness score. These measures are necessary, but they do not show whether a safety intervention removed a load-bearing…