SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs Via Stable Safety-Critical Expert Identification

Large language models based on Mixture-of-Experts have achieved substantial gains in efficiency and scalability, yet their architectural uniqueness introduces underexplored safety alignment challenges. Existing safety alignment strategies, predominantly designed for dense models, are ill-suited t...

L1ght Looks to Protect Internet Users from Toxic and Predatory Behavior

Cybersecurity has been regarded as a necessity for all computer users, especially today when data breaches and malware attacks have become rampant. However, one of the more overlooked aspects of cybersecurity is the prevention of other forms of cybercrime, such as the spread of harmful content an...