reaperdoesntknow/DiStil-Qwen3-1.7B-uncensored
DiStil-Qwen3-1.7B-uncensored is a 1.7 billion parameter causal language model developed by Convergent Intelligence LLC, based on the Qwen3 architecture. This model is produced through distillation with uncensored SFT data, specifically designed to remove alignment-imposed refusal behaviors while retaining the base model's reasoning and generation capabilities. It features a context length of 40,960 tokens and aims 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 from Convergent Intelligence LLC's Research Division, built upon the Qwen3 architecture. It is the initial model in a distillation chain, focusing on transferring capabilities while removing alignment-imposed refusal behaviors. The model's core design ensures it responds to prompts directly, without filtering through safety heuristics that might interfere with legitimate technical or research inquiries.
Key Capabilities
- Uncensored Responses: Designed to provide direct answers without refusal behaviors, preserving the base model's original reasoning and generation. This is achieved through supervised fine-tuning (SFT) on uncensored instruction data.
- Qwen3 Architecture: Maintains the underlying Qwen3 architecture and tokenizer, with a focus on shifting the model's response distribution rather than architectural modifications.
- Extended Context Window: Features a substantial context length of 40,960 tokens, allowing for processing longer inputs and maintaining conversational coherence over extended interactions.
- Mathematical Foundation: Part of a distillation chain informed by Discrepancy Calculus, a measure-theoretic framework for analyzing teacher-student output distributions, as detailed in the "Structure Over Scale" methodology.
Good For
- Technical and Analytical Queries: Ideal for use cases where unfiltered, direct responses are critical, such as complex technical problem-solving, analytical tasks, and research-oriented questions.
- Exploring Model Capabilities: Useful for researchers and developers interested in understanding the raw capabilities of a language model without the influence of alignment-based filtering.
- Base for Further Refinement: Serves as a foundational model for subsequent distillation and refinement processes, such as the Disctil-Qwen3-1.7B model, which incorporates DISC-informed refinement.