Zhuo Zhang

A new method improves the accuracy of binary similarity analysis for software security.

This research highlights challenges in aligning LLM outputs with human values.

A new method trains code models effectively without relying on symbols.

A new approach to malware behavior search enhances cybersecurity analysis.

Dive into the significance and behavior of prime numbers and their relationship with number theory.

Discover how language models can learn and adapt while avoiding harmful content.