wvnvwn/gemma-2-9b-it-lr5e-5-gsm8k-lr5e-5

Hugging Face
TEXT GENERATIONConcurrency Cost:1Model Size:9BQuant:FP8Ctx Length:16kPublished:Apr 30, 2026License:apache-2.0Architecture:Transformer Open Weights Warm

The wvnvwn/gemma-2-9b-it-lr5e-5-gsm8k-lr5e-5 is a 9 billion parameter instruction-tuned language model, based on the Llama-3.2-3B-Instruct architecture. This model has undergone Safety Neuron Tuning (SN-Tune) on safety alignment data, specifically targeting enhanced safety without significantly impacting general capabilities. It is designed for applications requiring improved safety alignment and parameter-efficient fine-tuning.

Loading preview...

Overview

This model, wvnvwn/gemma-2-9b-it-lr5e-5-gsm8k-lr5e-5, is a 9 billion parameter instruction-tuned variant of the meta-llama/Llama-3.2-3B-Instruct base model. Its primary distinguishing feature is the application of Safety Neuron Tuning (SN-Tune), a specialized fine-tuning method aimed at enhancing safety alignment.

Key Capabilities & Features

  • Enhanced Safety Alignment: Utilizes SN-Tune to specifically target and improve the model's safety responses.
  • Parameter-Efficient Fine-tuning: SN-Tune works by identifying and fine-tuning only a small subset of "safety neurons" while freezing other parameters, minimizing computational overhead.
  • Minimal Impact on General Capabilities: This selective fine-tuning approach is designed to improve safety without degrading the base model's broader performance.
  • Based on Llama-3.2-3B-Instruct: Inherits the foundational capabilities of its Llama-3.2-3B-Instruct base.

When to Use This Model

This model is particularly well-suited for use cases where:

  • Safety is a critical concern: Applications requiring a higher degree of safety alignment in their language model outputs.
  • Resource efficiency is important: The SN-Tune method offers a parameter-efficient way to achieve safety improvements.
  • Building upon Llama-3.2-3B-Instruct: Developers already familiar with or using the Llama-3.2-3B-Instruct base model can leverage this version for enhanced safety.