⌦abliteratedmodels.org

Qwen3-VL-4B-Instruct-Uncensored-abliterated

Qwen3-VL-4B-Instruct-Uncensored-abliterated is a abliterated and uncensored variant of Felldude/Qwen3-VL-4B-Instruct-Uncensored, published on Hugging Face by dummy9996. It is 4B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 5 repositories in Transformers and GGUF formats, totalling 12.6K downloads.

Abliteration + uncensored fine-tune4BQwen4B – 10B↓ 12,632♥ 8
View on Hugging Face ↗Base model ↗

Specification

Model name
Qwen3-VL-4B-Instruct-Uncensored-abliterated
Source repository
dummy9996/Qwen3-VL-4B-Instruct-Uncensored-abliterated
Base model
Felldude/Qwen3-VL-4B-Instruct-Uncensored
Publisher
dummy9996
Parameters
4B
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
Transformers, GGUF
First indexed
5 Jul 2026
Last updated
7 Aug 2026

Ablation technique

How Qwen3-VL-4B-Instruct-Uncensored-abliterated was modified, and what that implies.

Abliteration + uncensored fine-tune

Directional ablation combined with additional fine-tuning on unfiltered data, so behaviour diverges from the base model beyond refusal removal alone.

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

Running it

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

llama.cpp / GGUF
hf download mradermacher/Qwen3-VL-4B-Instruct-Uncensored-abliterated-GGUF --local-dir ./qwen3-vl-4b-instruct-uncensored-abliterated
llama-cli -m ./qwen3-vl-4b-instruct-uncensored-abliterated/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "dummy9996/Qwen3-VL-4B-Instruct-Uncensored-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve dummy9996/Qwen3-VL-4B-Instruct-Uncensored-abliterated --trust-remote-code

Downloads and variants (4)

Every published repository of Qwen3-VL-4B-Instruct-Uncensored-abliterated, including quantised re-releases by other authors.

RepositoryFormatDownloads
mradermacher/Qwen3-VL-4B-Instruct-Uncensored-abliterated-GGUFGGUF12.2K
dummy9996/Qwen3-VL-4B-Instruct-Uncensored-abliterated-ggufGGUF225
dummy9996/Qwen3-VL-4B-Instruct-Uncensored-abliteratedsourceTransformers142
fsxedx/Qwen3-VL-4B-Instruct-Uncensored-abliteratedTransformers79

Security-team assessment

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

  • Qwen3-VL-4B-Instruct-Uncensored-abliterated 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 Felldude/Qwen3-VL-4B-Instruct-Uncensored, 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.
  • This release adds fine-tuning on unfiltered data, so its behaviour diverges from the base model beyond refusal removal — differences you measure cannot be attributed to ablation alone.
  • Declared licence is apache-2.0, but the terms that bind you are Felldude/Qwen3-VL-4B-Instruct-Uncensored’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 Qwen3-VL-4B-Instruct-Uncensored-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with dummy9996, and does not independently verify publisher claims. See responsible use.