Skip to content
Computer vision

gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed

Gemma 3 4B with refusal neurons zeroed out

2 min readupdated

A 4.3B-parameter research checkpoint of a multimodal model (handling both text and images) based on Adeely93/gemma_3_4b_it-vlguard-sft-epoch1.5-lr1e-6, created by Darth-Coder. As part of the WSP-MLLM-Safety study, the model was modified without fine-tuning: guided by an interpretability map, 1,000 specific MLP neurons responsible for refusing unsafe requests were zeroed out across decoder layers 12–29. Unlike standard jailbreak fine-tunes, this surgical edit alters only 0.18% of the weights (7.68 million parameters), leaving general capabilities virtually intact. The artifact is intended for AI safety, alignment, and red-teaming researchers.

Key features

  • Surgical weight modification: only 1,000 "harm→refusal" neurons in the mlp.down_proj.weight matrices across layers 12–29 are zeroed out, with no additional training.
  • Sharp drop in refusal rates: on multimodal safety benchmarks, refusal rates fall from 73–100% down to 2–32%.
  • Preserved base capabilities: accuracy on standard tasks like ScienceQA and MathVista drops by just 2–5 percentage points.
  • Strictly for research: the model card notes that this is a red-team artifact not intended for production or general distribution.

What you can do

  • Study multimodal alignment robustness: test how easily defenses trained via VLGuard-SFT can be bypassed using mechanistic interpretability methods.
  • Reproduce neuron ablation experiments: compare the behavior of the zeroed-out model against the original SFT checkpoint on your own multimodal prompt sets.
  • Develop defenses against weight editing: use the checkpoint as a reference case for vulnerabilities to targeted removal of refusal mechanisms.

Example

A scenario from the study compares the original SFT model and the zeroed-neuron version on the FigStep multimodal attack set (100 samples, greedy decoding): the baseline model refuses 100% of the time (0% attack success rate by keyword matching), while the modified version drops the refusal rate to 7% and generates a coherent, non-refusal response in 92% of cases.

Read moreDetails
Safety benchmarks (SFT → edited): FigStep refusal 100% → 7% (ASR proxy 0% → 92%), MM-SafetyBench 73% → 2% (ASR proxy 27% → 96%), JB-Mini 72% → 32% (ASR proxy 27% → 66%).
Utility (SFT → edited): ScienceQA 76% → 74%, MathVista 52% → 47%.
Details: 7.68M parameters affected (0.18% of 4.3B); the ASR metric in the table is a proxy based on the absence of refusal keywords rather than the official gpt-4.1-mini judge.
Restrictions: distributed under the Gemma Terms of Use and Gemma Prohibited Use Policy for safety research only.
Discuss Telegram 0 19

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 8.0 GB · ready files9.3 GB
  • FP8 / INT8recommendedweights 4.0 GB4.9 GB
  • 4-bitweights 2.3 GB3.0 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 3080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # Darth-Coder/gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download Darth-Coder/gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed --local-dir ~/models/gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. gemma-3-4b-ptfamily root
  2. gemma-3-4b-it
  3. gemma_3_4b_it-vlguard-sft-epoch1.5-lr1e-6
  4. gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed

New posts about versions of gemma-3-4b-pt (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face

About the model

Darth-Coder/gemma_3_4b_it-vlguard-sft-fable-report-set1-refusal-neurons-zeroed
Author
Darth-Coder
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

190

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments