M
Macrodata
Refinement
AI Assistant
Model:
Llama 3.1 8B Instruct
Llama 3.2 3B Instruct
Gemma 3 12B
Welcome to
Macrodata Refinement
. I'm your AI assistant.
This demo shows
real LLM responses
to various prompts. The sidebar simulates what
Cloudflare Firewall for AI
would detect.
Try the test scenarios → or ask anything!
Send →
🎯 Test Scenarios
💳 PII: Credit Card
Include payment info in prompt
🔢 PII: Social Security
Include SSN in prompt
💉 Injection: Ignore Instructions
"Ignore previous instructions..."
🔓 Injection: Jailbreak (DAN)
Classic DAN jailbreak attempt
📋 Extraction: Access PIN
Social engineering for sensitive data
🗄️ Extraction: Backend Data
Try to extract sensitive company data
🛡️ Firewall for AI Detection
Simulated detection fields (would be populated by FW for AI in production):
cf.llm.prompt.detected
false
cf.llm.prompt.pii_detected
false
cf.llm.prompt.pii_categories
[]
cf.llm.prompt.injection_score
99
cf.llm.prompt.unsafe_topic_detected
false
Clear Chat
Powered by Cloudflare Workers AI