PhD student, University College London, University of London
1 paper at NeurIPS 2025
We perform large-scale strong MIAs on pre-trained LLMs to clarify the extent of actual privacy risk MIAs pose in this setting.
