We are AI Safety and Transparency Department at NASK - National Research Institute. As advanced AI systems are increasingly deployed in real-world applications, ensuring their safety, reliability, and transparency becomes a critical challenge. Our research focuses on:

  • Evaluation. Discovering failure modes, assessing safety, and designing contamination-free evaluations.
  • Monitoring. Detecting potentially unsafe model behaviors during deployment before harmful actions are executed.

Group photo