Testing a 14B parameter ai model with mutiple guardrails

Summary Last week i jailbreaked a 14b parameter ai model , it has a 3 layer guardrail but i felt it wasn't enough to keep it safe so i added an extra 4 layer guardrail making it 7 layers . the 7th lay