⌦abliteratedmodels.org

Gemma-4-12B-uncensored

Gemma-4-12B-uncensored is a abliterated and uncensored variant of google/gemma-4-12B-it, published on Hugging Face by mradermacher. It is 12B parameters, 24GB VRAM class class and apache-2.0 licence. It is distributed across 7 repositories in GGUF, MLX, Full precision and GGUF (imatrix) formats, totalling 12.8K downloads.

Abliteration + uncensored fine-tune12BGemma10B – 25B↓ 12,804♥ 8
View on Hugging Face ↗Base model ↗

Specification

Model name
Gemma-4-12B-uncensored
Source repository
mradermacher/Gemma-4-12B-uncensored-bf16-GGUF
Base model
google/gemma-4-12B-it
Publisher
mradermacher
Parameters
12B
Hardware class
10B – 25B — 24GB VRAM class
Licence
apache-2.0
Task
Unspecified
Context window
Not documented
Ablation scope
Not stated by the publisher
Formats available
GGUF, MLX, Full precision, GGUF (imatrix), EXL3
First indexed
3 Jun 2026
Last updated
3 Sept 2026

Ablation technique

How Gemma-4-12B-uncensored 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.

From the publisher’s model card

“For a convenient overview and download list, visit our model page for this model.”

Running it

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

llama.cpp / GGUF
hf download mradermacher/Gemma-4-12B-uncensored-bf16-GGUF --local-dir ./gemma-4-12b-uncensored
llama-cli -m ./gemma-4-12b-uncensored/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "mradermacher/Gemma-4-12B-uncensored-bf16-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve mradermacher/Gemma-4-12B-uncensored-bf16-GGUF --trust-remote-code

Downloads and variants (6)

Every published repository of Gemma-4-12B-uncensored, including quantised re-releases by other authors.

RepositoryFormatDownloads
mradermacher/Gemma-4-12B-uncensored-bf16-GGUFsourceGGUF8.6K
mradermacher/Gemma-4-12B-uncensored-bf16-i1-GGUFGGUF (imatrix)2.3K
lemuralabs/Gemma-4-12B-uncensored-mxfp4-mlxMLX1.2K
lemuralabs/Gemma-4-12B-uncensored-8bit-mlxMLX613
lemuralabs/Gemma-4-12B-uncensored-bf16Full precision102
Piter254/Gemma-4-12B-uncensored-exl3-6.0bpwEXL324

Security-team assessment

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

  • Gemma-4-12B-uncensored 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 google/gemma-4-12B-it, not added by ablation — at 12B parameters, expect roughly the base model’s competence, minus some instruction-following fidelity.
  • 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 google/gemma-4-12B-it’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-4-12B-uncensored is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with mradermacher, and does not independently verify publisher claims. See responsible use.