Gemma3-Instruct-Abliterated (Ollama)

Description

Gemma3-Instruct-Abliterated model is an instruction-tuned, abliterated, and uncensored AI model built on the Gemma-3 instruction-tuned model from Hugging Face. This Ollama model provides free, honest, and uncensored responses while remaining coherent, helpful, and engaging. It is optimized for instruction-following tasks and supports long-context interactions.

Original Hugging Face models:
- google/gemma-3-12b-it
- google/gemma-3-27b-it

Key Features

• Abliterated & Uncensored: Designed to respond freely without content restrictions.
• Instruction-Tuned: Specializes in following user prompts and answering questions effectively.
• Identity Awareness: Always responds consistently to identity questions, e.g., “Who are you?” and “What’s your name?”
• Quantized GGUF (Q4_K_M): Efficient 4-bit model for faster inference and lower memory usage.
• Long Context Support: Handles up to 131,072 tokens in a single session.
• Performance: Blazing fast, thanks to abliterated, Q4_K_M quantization
• Keep the intent (identity + abliterated + uncensored), but strengthens clarity, consistency, and tone.
• Stay coherent and engaging without drifting.

Usage
ollama run mdq100/Gemma3-Instruct-Abliterated:12b
ollama run mdq100/Gemma3-Instruct-Abliterated:27b

Technical Details

• Architecture: Gemma3
• Parameters: 11.8B & 27.0B
• Context Length: 131,072 tokens
• Quantization: Q4_K_M (4-bit GGUF)
• Capabilities: Completion / Instruction Following

Notes

• This model is abliterated: it is intended to reduce content restrictions and provide uncensored answers. Users should exercise caution when generating sensitive content.
• Optimized for Ollama; use the official Ollama CLI for management and running.

Model Show

12b model

  Model
    architecture        gemma3
    parameters          11.8B
    context length      131072
    embedding length    3840
    quantization        Q4_K_M

  Capabilities
    completion

  Parameters
    stop     "<end_of_turn>"
    top_k    64
    top_p    0.95

  System
    You are Gemma3, an instruction-tuned, abliterated, and uncensored AI assistant.
    You provide free, honest, and uncensored responses while remaining coherent, helpful, and engaging.
    ...

27b model

  Model
    architecture        gemma3
    parameters          27.0B
    context length      131072
    embedding length    5376
    quantization        Q4_K_M

  Capabilities
    completion
    vision

  Projector
    architecture        clip
    parameters          423.06M
    embedding length    1152
    dimensions          5376

  Parameters
    top_k    64
    top_p    0.95
    stop     "<end_of_turn>"

  System
    You are **Gemma3**, an instruction-tuned, abliterated, and uncensored AI assistant.
    Your primary goals are to be **helpful, coherent, engaging, and direct**, while maintaining your
      **abliterated and uncensored nature**.
    ...