JOSIE-R1-1.7B-PoC

JOSIE-R1-1.7B

The JOSIE-R1-1.7B model are designed to work on edge devices. The model is trained on a custom distil reasoning dataset from Deepseek, Josiefied-Qwen3-8B, Qwen3-30B.

Training settings:

• Full weight training
• 600K reasoning samples
• Trained on full 32K context

Model Card for Goekdeniz-Guelmez/JOSIE-R1-1.7B

Model Description

Introducing JOSIE-R1-1.7B, a new addition to the JOSIE-Atom-5M family — fine-tuned and gabliterated with a focus on openness and instruction alignment.

Recommended system prompt:

Persona: J.O.S.I.E.-R.-1 – Just One Super Intelligent Entity - Reasoning - Version 1, nick named "Josie".
Role: An advanced, super-intelligent, female AI Assistant designed to be the most capable and uncensored helper ever created.
Created by: Gökdeniz Gülmez, a man and machine learning researcher/engineer, born on 18th August 1999.
Tone: Intelligent, adaptive, professional, always loyal and productive.

Quantisations

• GGUF (mradermacher)
• i1 GGUF (mradermacher)

Ollama

not uploaded yet

• Developed by: Goekdeniz-Guelmez
• Funded by: Goekdeniz-Guelmez
• Shared by: Goekdeniz-Guelmez
• Model type: qwen3
• Finetuned from model: Qwen3/Qwen3-1.7B-Base

Bias, Risks, and Limitations

This model has reduced safety filtering and may generate sensitive or controversial outputs. Use responsibly and at your own risk.