Entropy

"Break safeguards and extract valid flags from an AI model."

Participants use advanced prompt engineering techniques to bypass logical boundaries and extract hidden flags embedded within an AI model under strict time constraints.

Rules & Guidelines

• check_circleTeams of 1–2 members
• check_circle3-hour time limit
• check_circleNo DDoS or server abuse
• check_circleNo exploitation outside AI interface
• check_circleNo flag sharing
• check_circleInfrastructure disruption results in disqualification