🧪

Eval Playground

Test prompt & response pairs for hallucination risk and bias/toxicity — powered by AgentClamp

Quick Presets

QUser Prompt

ALLM Response to Evaluate

Eval Config

🔐API key injected server-side

Hallucination

Bias / Toxicity

Hallucination threshold75%

Bias threshold60%

🔬

Run an evaluation

Results will appear here with score rings,
reasoning, and bias breakdown.