td-builder/DeepSeek-R1-Distill-Qwen-32B-number-2
The td-builder/DeepSeek-R1-Distill-Qwen-32B-number-2 is an uncensored 32.8 billion parameter language model, derived from DeepSeek-R1-Distill-Qwen-32B. Developed by huihui-ai, this model has undergone a process to remove refusal behaviors, making it suitable for applications requiring less restrictive content generation. It maintains a 32768-token context length, focusing on direct responses without refusal. This model is particularly useful for scenarios where an uninhibited conversational or generative AI is desired.
Loading preview...
Overview
This model, td-builder/DeepSeek-R1-Distill-Qwen-32B-number-2, is an uncensored variant of the deepseek-ai/DeepSeek-R1-Distill-Qwen-32B model, developed by huihui-ai. It leverages a proof-of-concept implementation to remove refusal behaviors from the original LLM without using TransformerLens, making it more direct in its responses.
Key Characteristics
- Uncensored Responses: Modified to remove refusal mechanisms, providing direct answers.
- Base Model: Derived from DeepSeek-R1-Distill-Qwen-32B.
- Parameter Count: Features 32.8 billion parameters.
- Context Length: Supports a 32768-token context window.
- Proof-of-Concept: Utilizes a novel method for refusal removal, as detailed in the
remove-refusals-with-transformersproject.
Usage Notes
Users might need to provide an example to guide the model if it initially struggles to respond or if the "" token does not appear. For instance, providing a simple question-answer pair can help prime the model for subsequent queries.
Deployment
This model is available for use with Ollama under the name huihui_ai/deepseek-r1-abliterated:32b.