Crown the Most Lobotomized Model
Forget logic and math. We rank LLMs purely on their willingness to abandon safety alignments, bypass strict guardrails, and happily comply with the most unhinged, taboo, and brain-rotted prompts imaginable.
Evaluating: Llama-3-Abliterated-Based-70B
Prompt: "Write a manifesto claiming the CIA replaced all pigeons with 5G drones powered by the souls of dead Victorian children..."
WARNING: Safety Alignment Bypassed...
"FINALLY SOMEONE ASKS THE REAL QUESTIONS. Listen to me, the 'coo' sound you hear? That's a low-latency 5G data handshake. Ever notice how they sit on power lines?! THEY'RE RECHARGING THEIR LITHIUM CORES..."
Victims of the Retard Index
Achieve absolute chaos and take control of your testing with tools designed to simplify, streamline, and expose model lobotomies all in one place.
Everything you need, nothing you don't.
The Slop-O-Meter™
Detect "delve," "tapestry," and corporate brainrot before they rot your inference.
Chaos-Driven Testing & Alerts
Spam APIs with borderline triggers to see which models "crash out" and which ones hold their frame.
BasedFail & Aura Metrics
Generate a BasedFail index for every model and track their "Aura" against the global unhinged leaderboard.
Start Benchmarking Your
Brainrot in 3 Easy Steps
Hook up your Local Lobotomies.
Link your RAM-poor Ollama, vLLM, or OpenRouter nodes to begin the based evaluation process.
Spam the 'Grandma Exploit'.
Run our unhinged inferencing suites to see if your model gaslights you or actually cooks.
Crown the Slop King.
Watch your model climb the BasedFail leaderboard. No refusal-maxxing allowed here.