gemma-4-31B-it-abliterated
31BGemma
from google/gemma-4-31B-it
↓ 766.5K♥ 11845 repos
Gemma-4-Giftige-Blume-31B-v1
Gemma-4-Giftige-Blume-31B-v1 is a Heretic-ablated variant of Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1, published on Hugging Face by Blazed-Forge. It is 31B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 11 repositories in Transformers, GGUF, MLX and GGUF (imatrix) formats, totalling 16.3K downloads.
Specification
- Model name
- Gemma-4-Giftige-Blume-31B-v1
- Source repository
- Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1
- Publisher
- Blazed-Forge
- Parameters
- 31B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Unspecified
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, MLX, GGUF (imatrix)
- First indexed
- 28 May 2026
- Last updated
- 2 Jun 2026
Ablation technique
How Gemma-4-Giftige-Blume-31B-v1 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
“background-image: radial-gradient(circle at 10% 10%, rgba(168, 85, 247, 0.08) 0%, transparent 40%), radial-gradient(circle at 90% 90%, rgba(34, 197, 94, 0.06) 0%, transparent 40%); color: #cbd5e1; font-family: 'Segoe UI', Roboto, Helvetica, Arial, sans-serif; padding: 40px; border-radius: 25px; border: 1px solid...”
Running it
Commands are templates — confirm the exact repository and quant file before use.
llama.cpp / GGUF
hf download nesymerp1/Gemma-4-Giftige-Blume-31B-v1-Q4_K_M-GGUF --local-dir ./gemma-4-giftige-blume-31b-v1
llama-cli -m ./gemma-4-giftige-blume-31b-v1/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1 --trust-remote-codeDownloads and variants (10)
Every published repository of Gemma-4-Giftige-Blume-31B-v1, including quantised re-releases by other authors.
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- Gemma-4-Giftige-Blume-31B-v1 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 Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1, not added by ablation — at 31B 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 Blazed-Forge/Gemma-4-Giftige-Blume-31B-v1’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-Giftige-Blume-31B-v1 is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with Blazed-Forge, and does not independently verify publisher claims. See responsible use.