QExplorer leverages the power of Large Language Models to automatically extract effective search queries for identifying disguised toxic content online.

• Addresses the challenge of detecting harmful content that is intentionally obscured
• Utilizes generative LLM capabilities to create targeted queries for content exploration
• Improves security systems' ability to discover similar toxic content
• Demonstrates practical application of LLMs for online safety and content moderation

Security Impact: This research provides security teams with more effective tools to identify and mitigate harmful content that traditional keyword-based systems might miss, enhancing online safety measures.

Original Paper: QExplorer: Large Language Model Based Query Extraction for Toxic Content Exploration