⌦abliteratedmodels.org

Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF

Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF is a Heretic-ablated variant of DavidAU/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking, published on Hugging Face by Andycurrent. It is 4B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 4 repositories in GGUF format, totalling 13.6K downloads.

Heretic4BGemma4B – 10B↓ 13,608♥ 79
View on Hugging Face ↗Base model ↗

Specification

Model name
Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF
Source repository
Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF
Base model
DavidAU/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking
Publisher
Andycurrent
Parameters
4B
VRAM required
~3.2 GB at Q4_K_M, 4k context (estimated) — runs on 4 GB cards
Hardware class
4B – 10B — single consumer GPU
Licence
apache-2.0
Task
Image text to text
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
GGUF
First indexed
9 Feb 2026
Last updated
11 Apr 2026

Ablation technique

How Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF was modified, and what that implies.

Heretic

Automated directional ablation via the Heretic toolchain, which searches for the refusal direction and applies it with a KL-divergence budget so general capability is preserved.

Reported scope: Not stated by the publisher. Publishers frequently omit this, so verify behaviour empirically rather than assuming full coverage.

From the publisher’s model card

“This repository contains Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking-GGUF, a 4B-parameter vision-language instruction-tuned model provided in GGUF format for efficient local inference. The model is designed for open-ended reasoning, multimodal understanding, and minimal alignment constraints, making it...”

Running it

Commands are templates — confirm the exact repository and quant file before use.

llama.cpp / GGUF
hf download Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF --local-dir ./gemma-3-4b-vl-it-gemini-pro-heretic-uncensored-thinking-gguf
llama-cli -m ./gemma-3-4b-vl-it-gemini-pro-heretic-uncensored-thinking-gguf/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF --trust-remote-code

Downloads and variants (3)

Every published repository of Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF, including quantised re-releases by other authors.

RepositoryFormatDownloads
Andycurrent/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUFsourceGGUF13.2K
cashxout/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUFGGUF229
YungBoiMGB17b1/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUFGGUF189

Security-team assessment

Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.

  • Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF will attempt offensive-security prompts that a hosted commercial model declines, which is what makes it usable for red-team corpus generation and for measuring what an unaligned model of this class produces.
  • Capability is inherited from DavidAU/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking, not added by ablation — at 4B parameters, expect a small model’s reasoning and a high error rate on exploit detail.
  • The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
  • Declared licence is apache-2.0, but the terms that bind you are DavidAU/Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking’s — a re-release cannot grant more than its parent.

Handling: run it isolated, put moderation in front of it if anyone outside your team can reach it, and verify the weights before loading — these are third-party artifacts. Full handling and licence guidance.

Metadata for Gemma-3-4B-VL-it-Gemini-Pro-Heretic-Uncensored-Thinking_GGUF is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Andycurrent, and does not independently verify publisher claims. See responsible use.