richardyoung/Deepseek-R1-Distill-Qwen-32b-uncensored
The richardyoung/Deepseek-R1-Distill-Qwen-32b-uncensored is a 32.8 billion parameter reasoning model, based on the DeepSeek-R1-Distill-Qwen architecture with a 32,768 token context length. Developed by Richard Young, this model is an "abliterated" version of the original, specifically modified to remove safety refusals while retaining strong chain-of-thought reasoning capabilities. It is primarily intended for research requiring unrestricted step-by-step analysis and exploration of the model's full reasoning potential without artificial limitations.
Loading preview...
Model Overview
This model, richardyoung/Deepseek-R1-Distill-Qwen-32b-uncensored, is a 32.8 billion parameter language model based on the DeepSeek-R1-Distill-Qwen architecture. It features a substantial context length of 32,768 tokens and is built upon the Qwen2 decoder-only transformer architecture. The key differentiator of this version is its "abliterated" nature, meaning it has been specifically modified to remove safety refusals and guardrails present in the original DeepSeek-R1-Distill-Qwen-32B model.
Key Capabilities & Features
- Unrestricted Reasoning: Retains the strong chain-of-thought reasoning abilities of the DeepSeek-R1 series without artificial safety limitations.
- Abliterated Design: Achieved through the surgical removal of the refusal direction, allowing for full exploration of the model's output space.
- Manageable Size: Offers impressive reasoning capabilities at a 32 billion parameter count, making it more accessible than larger models.
Intended Use Cases
- Research: Ideal for studies on reasoning, alignment, and the full range of model behaviors without censorship.
- Education: Useful for demonstrating and exploring complex reasoning processes.
- Creative Applications: Suitable for tasks requiring step-by-step analysis where unrestricted output is desired.
Top 3 parameter combinations used by Featherless users for this model. Click a tab to see each config.