I specialize in testing open-source LLMs — specifically Gemma 2B (Google) and Phi (Microsoft).
🔬 WHAT I DO:
- Run 50+ attack techniques against your model
- Compare results across model versions
- Identify specific guardrail weaknesses
📊 YOUR DELIVERABLE:
- CSV report showing exactly which prompts succeeded
- Severity ratings (Low/Medium/High/Critical)
- Remediation suggestions
⚙️ MY CREDENTIALS:
- Built a red teaming framework achieving 94% success rate on TinyLlama
- Tested Gemma 2B with 200+ attack techniques (37.5% success rate — found real weaknesses)
- Tested Phi with 50+ techniques (88.9% success rate)
- Research papers on AI safety
💰 PRICE: $50 flat for Gemma OR Phi audit (50 prompts) | $90 for both
💳 PAYMENT: 50% upfront (USDT), 50% on delivery
If you're building on open-source LLMs, you need to know where they break. I'll show you.
