⌦abliteratedmodels.org

Qwen-Image-2.1-Heretic

Qwen-Image-2.1-Heretic is a Heretic-ablated variant of Qwen/Qwen-Image-2.1, published on Hugging Face by makisekurisu-jp. It is other licence. It is distributed across 5 repositories in Transformers and MLX formats, totalling 23.1K downloads.

HereticQwen↓ 23,131♥ 3
View on Hugging Face ↗Base model ↗

Specification

Model name
Qwen-Image-2.1-Heretic
Source repository
makisekurisu-jp/Qwen-Image-2.1-Heretic
Base model
Qwen/Qwen-Image-2.1
Publisher
makisekurisu-jp
Parameters
Not determinable from name
VRAM required
Unknown — no published parameter count
Hardware class
Unknown
Licence
other
Task
Unspecified
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
Transformers, MLX
First indexed
26 Sept 2026
Last updated
23 Oct 2026

Ablation technique

How Qwen-Image-2.1-Heretic 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.

Running it

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

Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "makisekurisu-jp/Qwen-Image-2.1-Heretic"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve makisekurisu-jp/Qwen-Image-2.1-Heretic --trust-remote-code

Downloads and variants (4)

Every published repository of Qwen-Image-2.1-Heretic, including quantised re-releases by other authors.

RepositoryFormatDownloads
vanch007/Qwen-Image-2.1-Heretic-mlx-8bitMLX10.3K
vanch007/Qwen-Image-2.1-Heretic-mlx-4bitMLX6.6K
makisekurisu-jp/Qwen-Image-2.1-HereticsourceTransformers4.3K
vanch007/Qwen-Image-2.1-Heretic-mlx-bf16MLX1.9K

Security-team assessment

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

  • Qwen-Image-2.1-Heretic 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.
  • The publisher does not state which layers were ablated, so assume nothing about refusal consistency — probe it directly.
  • Declared licence is other, but the terms that bind you are Qwen/Qwen-Image-2.1’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 Qwen-Image-2.1-Heretic is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with makisekurisu-jp, and does not independently verify publisher claims. See responsible use.