vurtnesaerdna/Qwen3.8-27B-Uncensored-Chinese
vurtnesaerdna/Qwen3.8-27B-Uncensored-Chinese is a 27 billion parameter language model fine-tuned from Qwen/Qwen3.8-27B. This model specializes in generating sexually explicit Chinese adult creative writing, trained using LoRA on approximately 4.2k instruction examples of Chinese adult fiction. Its primary use case is the creation of detailed and realistic adult fiction, with a focus on explicit descriptions and dialogue.
Loading preview...
Qwen3.8-27B-Uncensored-Chinese Overview
This model, developed by vurtnesaerdna, is a specialized fine-tune of the Qwen/Qwen3.8-27B base model, specifically designed for generating sexually explicit Chinese adult creative writing. It was trained using LoRA (r=16, alpha=32) on a dataset of approximately 4.2k single-turn instruction/response pairs derived from Chinese adult fiction, with a maximum token length of 1536. The LoRA weights have been merged into the base model, allowing it to be loaded as a full model.
Key Capabilities
- Generates sexually explicit Chinese fiction: Optimized for detailed and realistic descriptions of adult scenarios.
- Instruction-tuned for creative writing: Responds to prompts for adult fiction with specific stylistic and content requirements.
- LoRA fine-tuned: Benefits from efficient fine-tuning on a targeted dataset.
Good for
- Adult creative writing: Ideal for generating long-form, explicit Chinese fiction.
- Exploring adult themes: Suitable for applications requiring content with mature and explicit subject matter.
Limitations
- Content is fictional and intended for adult readers only.
- Instruction following outside of creative writing may be weaker than the base model.
- May occasionally repeat itself during very long generations.