After KAUST Academy
Nasser attended a program at Tsinghua University in Beijing in September 2026, with lectures, technology company visits and an AI exhibition. Nasser's team had won the Best Poster Award in August 2026 for research on AI safety in diffusion language models.
Summer project
Artificial Intelligence Specialization · Cohort 2026
Mechanistic Analysis of Template-Injection Jailbreaks in Diffusion Language Models
Our project investigates how jailbreak attacks affect diffusion language models in healthcare. We study whether unsafe template injections leave detectable internal signatures and whether these signals can be used to steer the model back toward safer responses. Our findings show that harmful injections can be detected reliably, but correcting the resulting behavior remains more challenging.
With KAUST Academy
- Cohort 2026
Artificial Intelligence Specialization
University of Toronto · Artificial Intelligence Summer Program
Want your entry updated or removed? Contact us.