gemma-4-26B-A4B-it-qat-uncensored-heretic
26BGemmaheretic
from llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic
↓ 24K♥ 274 repos
gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov
gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov is a Heretic-ablated variant of llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic, published on Hugging Face by Wondernutts. It is 26B total parameters (mixture-of-experts), 48GB VRAM class class and apache-2.0 licence. It is distributed across 2 repositories in Transformers format, totalling 7.5K downloads.
Specification
- Model name
- gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov
- Publisher
- Wondernutts
- Parameters
- 26B (mixture-of-experts)
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Unspecified
- Context window
- 131K tokens (as documented)
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers
- First indexed
- 15 Jun 2026
- Last updated
- 6 Sept 2026
Ablation technique
How gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov 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
“This checkpoint: the headline benchmark used the local LUT131K Heretic deployment of this model. The release text/recall/cache gates passed; see the report for the exact configuration and limitations.”
Running it
Commands are templates — confirm the exact repository and quant file before use.
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "Wondernutts/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-int4-ov"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve Wondernutts/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-int4-ov --trust-remote-codeDownloads and variants (1)
The only published repository of gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov.
| Repository | Format | Downloads |
|---|---|---|
| Wondernutts/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-int4-ovsource | Transformers | 7.5K |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- gemma-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov 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 llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic, not added by ablation — at 26B 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 apache-2.0, but the terms that bind you are llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic’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-4-26B-A4B-it-qat-unquantized-uncensored-heretic-ov is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Wondernutts, and does not independently verify publisher claims. See responsible use.