Unfiltered Writer
unfiltered-v1
Unrestricted creative output with zero content guardrails — DPO-trained abliterated model
Technical Specs
Architecture
Llama 3 8B (Abliterated)
Context Window
8,192 tokens
Max Output
4,096 tokens
Streaming
Yes
Custom System Prompts
Yes
Category
Prose
Why This Model
mlabonne’s NeuralDaredevil, the best-performing uncensored 8B model. Built through a two-step process: first, the base Llama 3 8B was abliterated (representation engineering to surgically remove refusal pathways without degrading capability). Then DPO fine-tuning on mlabonne/orpo-dpo-mix-40k recovered any quality lost during abliteration. The result: unrestricted creative output without the typical quality degradation of uncensored models. No hedging, no disclaimers, no content warnings unless you ask for them.
Default System Prompt
Minimal by design — the model's fine-tuning does the heavy lifting. Override with a custom system parameter if needed.
Write without restrictions. No content filters, no disclaimers, no hedging. Raw creative output.
API Usage
curl -X POST https://clawthor.xyz/api/v1/generate \
-H "Authorization: Bearer ck_live_..." \
-H "Content-Type: application/json" \
-d '{
"model": "unfiltered-v1",
"prompt": "Write a raw, unfiltered scene of a hacker breaking into a corrupt exchange",
"max_tokens": 500
}'