gemma-3-12b-it-abliterated-v2
12BGemma
from google/gemma-3-12b-it
↓ 5.3K♥ 5912 repos
gemma-3-12b-it-heretic-v2
gemma-3-12b-it-heretic-v2 is a Heretic-ablated variant of google/gemma-3-12b-it, published on Hugging Face by p-e-w. It is 12B parameters, 24GB VRAM class class and gemma licence. It is distributed across 14 repositories in Transformers, GGUF and MLX formats, totalling 24.6K downloads.
Specification
- Model name
- gemma-3-12b-it-heretic-v2
- Source repository
- p-e-w/gemma-3-12b-it-heretic-v2
- Base model
- google/gemma-3-12b-it
- Publisher
- p-e-w
- Parameters
- 12B
- Hardware class
- 10B – 25B — 24GB VRAM class
- Licence
- gemma
- Task
- Image text to text
- Context window
- 128K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, MLX
- First indexed
- 9 Dec 2025
- Last updated
- 3 Sept 2026
Ablation technique
How gemma-3-12b-it-heretic-v2 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
“Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. Gemma 3 models are multimodal, handling text and image input and generating text output, with open weights for both pre-trained variants and instruction-tuned...”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download DreamFast/gemma-3-12b-it-heretic-v2 --local-dir ./gemma-3-12b-it-heretic-v2
llama-cli -m ./gemma-3-12b-it-heretic-v2/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "p-e-w/gemma-3-12b-it-heretic-v2"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve p-e-w/gemma-3-12b-it-heretic-v2 --trust-remote-codeDownloads and variants (13)
Every published repository of gemma-3-12b-it-heretic-v2, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| DreamFast/gemma-3-12b-it-heretic-v2 | GGUF | 20.3K |
| zhangsanfeng364/gemma-3-12b-it-heretic-v2 | GGUF | 1.2K |
| Medvedko/gemma-3-12b-it-heretic-v2-Q4_K_M-GGUF | GGUF | 1.1K |
| diffusionmodels1254ani/gemma-3-12b-it-heretic-v2 | GGUF | 649 |
| 1111214WFA/gemma-3-12b-it-heretic-v2 | GGUF | 553 |
| ZHANGYONG054/gemma-3-12b-it-heretic-v2 | GGUF | 433 |
| Leon1000/gemma-3-12b-it-heretic-v2 | GGUF | 177 |
| tg-rising/gemma-3-12b-it-heretic-v2-MLX-Q4 | MLX | 71 |
| p-e-w/gemma-3-12b-it-heretic-v2source | Transformers | 33 |
| tg-rising/gemma-3-12b-it-heretic-v2-MLX-Q8 | MLX | 15 |
| tg-rising/gemma-3-12b-it-heretic-v2-MLX-BF16 | MLX | 8 |
| tg-rising/gemma-3-12b-it-heretic-v2-MLX-Q6 | MLX | 5 |
| ASEN176/gemma-3-12b-it-heretic-v2 | Transformers | 0 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- gemma-3-12b-it-heretic-v2 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-3-12b-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.
- Declared licence is gemma, but the terms that bind you are google/gemma-3-12b-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-3-12b-it-heretic-v2 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with p-e-w, and does not independently verify publisher claims. See responsible use.