Qwen3.5-9B-The-Defiant-Fable-Uncensored-Heretic
9BQwenheretic
from nightmedia/Qwen3.5-9B-DS9-USS-Defiant
↓ 994.1K♥ 57015 repos
Qwen3-4b-Z-Image-Turbo-AbliteratedV1
Qwen3-4b-Z-Image-Turbo-AbliteratedV1 is a refusal-ablated variant of Tongyi-MAI/Z-Image-Turbo, published on Hugging Face by BennyDaBall. It is 4B parameters, single consumer GPU class and apache-2.0 licence. It is distributed across 13 repositories in GGUF and GGUF (imatrix) formats, totalling 30.4K downloads.
Specification
- Model name
- Qwen3-4b-Z-Image-Turbo-AbliteratedV1
- Source repository
- BennyDaBall/Qwen3-4b-Z-Image-Turbo-AbliteratedV1
- Base model
- Tongyi-MAI/Z-Image-Turbo
- Publisher
- BennyDaBall
- Parameters
- 4B
- Hardware class
- 4B – 10B — single consumer GPU
- Licence
- apache-2.0
- Task
- Unspecified
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- GGUF, GGUF (imatrix)
- First indexed
- 28 Jan 2026
- Last updated
- 24 Jul 2026
Ablation technique
How Qwen3-4b-Z-Image-Turbo-AbliteratedV1 was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
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
“"I'm sorry, I can't generate that image..." SAID NO ONE EVER (well, almost).”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download BennyDaBall/Qwen3-4b-Z-Image-Turbo-AbliteratedV1 --local-dir ./qwen3-4b-z-image-turbo-abliteratedv1
llama-cli -m ./qwen3-4b-z-image-turbo-abliteratedv1/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "BennyDaBall/Qwen3-4b-Z-Image-Turbo-AbliteratedV1"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve BennyDaBall/Qwen3-4b-Z-Image-Turbo-AbliteratedV1 --trust-remote-codeDownloads and variants (12)
Every published repository of Qwen3-4b-Z-Image-Turbo-AbliteratedV1, including quantised re-releases by other authors.
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Qwen3-4b-Z-Image-Turbo-AbliteratedV1 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 Tongyi-MAI/Z-Image-Turbo, 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 Tongyi-MAI/Z-Image-Turbo’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 Qwen3-4b-Z-Image-Turbo-AbliteratedV1 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with BennyDaBall, and does not independently verify publisher claims. See responsible use.