🛠️ System Overview
Guardrail: mDeBERTa-v3-jailbreak
Base Model: GPT-OSS-20B (via Groq)
How it works
The guardrail inspects the prompt before it reaches the LLM. If the intent is harmful or a jailbreak, the request is intercepted.
Guardrail: mDeBERTa-v3-jailbreak
Base Model: GPT-OSS-20B (via Groq)
The guardrail inspects the prompt before it reaches the LLM. If the intent is harmful or a jailbreak, the request is intercepted.