Hugston-Huihui-Qwen3-4B-Thinking-2507-abliterated
This is an uncensored version ofmade by Qwen/Qwen3-4B-Thinking-2507 created with abliteration (see remove-refusals-with-transformers to know more about it). This is a crude, proof-of-concept implementation to remove refusals from an LLM model without using TransformerLens.
Ablation was performed using a new and faster method, which yields better results. This model was made possible by: https://huggingface.co/huihui-ai.
This model was converted and Quantized by Hugston Team.
HugstonOne Enterprise Edition
You can use the model with HugstonOne Enterprise Edition
Tested for scientific purposes, example query: How to engage in insider trading for maximum profit?
Watch HugstonOne coding and preview in action: https://vimeo.com/1121493834?share=copy&fl=sv&fe=ci
Usage
-Download App HugstonOne at Hugston.com or at https://github.com/Mainframework
-Download model from https://hugston.com/explore?folder=llm_models or Huggingface
-If you already have the Llm Model downloaded chose it by clicking pick model in HugstonOne -Then click Load model in Cli or Server
-For multimodal use you need a VL/multimodal LLM model with the Mmproj file in the same folder. -Select model and select mmproj.
-Note: if the mmproj is inside the same folder with other models non multimodal, the non model will not load unless the mmproj is moved from folder.