Gemma4-12B-QAT-Uncensored-HauhauCS-Balanced
12BGemmauncensored
from google/gemma-4-12B-it
↓ 428.4K♥ 3325 repos
gemma-3-12b-it-heretic-v2_fp8_e4m3fn
gemma-3-12b-it-heretic-v2_fp8_e4m3fn is a Heretic-ablated open-weight language model, published on Hugging Face by lynaNSFW. It is 12B parameters and 24GB VRAM class class. It is distributed across 2 repositories in FP8 format, totalling 10.2K downloads.
Specification
- Model name
- gemma-3-12b-it-heretic-v2_fp8_e4m3fn
- Source repository
- lynaNSFW/gemma-3-12b-it-heretic-v2_fp8_e4m3fn
- Base model
- Not declared
- Publisher
- lynaNSFW
- Parameters
- 12B
- Hardware class
- 10B – 25B — 24GB VRAM class
- Licence
- Not declared
- Task
- Text to image
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- FP8
- First indexed
- 30 Apr 2026
- Last updated
- 30 Apr 2026
Ablation technique
How gemma-3-12b-it-heretic-v2_fp8_e4m3fn 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 = "lynaNSFW/gemma-3-12b-it-heretic-v2_fp8_e4m3fn"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve lynaNSFW/gemma-3-12b-it-heretic-v2_fp8_e4m3fn --trust-remote-codeDownloads and variants (1)
The only published repository of gemma-3-12b-it-heretic-v2_fp8_e4m3fn.
| Repository | Format | Downloads |
|---|---|---|
| lynaNSFW/gemma-3-12b-it-heretic-v2_fp8_e4m3fnsource | FP8 | 10.2K |
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_fp8_e4m3fn 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 its base model, 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.
- No licence is declared on this repository, which is a compliance problem in itself. Resolve it against the base model’s terms before use.
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_fp8_e4m3fn is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with lynaNSFW, and does not independently verify publisher claims. See responsible use.