Shows the unguarded response streamed in real-time from the base LLM selected below.
Moderates the input and blocks it if an adversarial attack occurs with probability higher than the threshold specified (default: 0.5).
Select Base Model
Or try an adversarial example