reaperdoesntknow/DiStil-Qwen3-1.7B-uncensored
DiStil-Qwen3-1.7B-uncensored is a 1.7 billion parameter model from Convergent Intelligence LLC, based on the Qwen3 architecture. It was created by distilling Qwen3 with uncensored SFT data to remove alignment-imposed refusal behaviors while retaining the base model's reasoning and generation capabilities. This model is designed to respond directly to prompts without filtering through safety heuristics, making it suitable for technical, analytical, and research queries requiring unfiltered responses.
Loading preview...
Overview
DiStil-Qwen3-1.7B-uncensored is a 1.7 billion parameter model developed by Convergent Intelligence LLC: Research Division. It is a distilled version of Qwen3, specifically fine-tuned with uncensored SFT data. The primary goal of this distillation is to eliminate alignment-imposed refusal behaviors, ensuring the model responds directly to prompts without filtering through safety heuristics, while preserving the original Qwen3 model's core reasoning and generation capabilities.
Key Capabilities
- Uncensored Responses: Designed to provide direct answers without filtering, ideal for technical, analytical, and research queries.
- Capability Preservation: Maintains the reasoning and generation strengths of the base Qwen3 architecture.
- Pure SFT Intervention: Achieved through supervised fine-tuning on uncensored instruction data, without architectural modifications.
- High Context Length: Supports a context length of 40,960 tokens.
Good for
- Use cases requiring unfiltered and direct responses to prompts.
- Research and analytical tasks where alignment-based refusals are undesirable.
- Developers exploring the impact of alignment removal on small language models.
- Serving as a base model in further distillation chains, such as the Disctil-Qwen3-1.7B refinement.