gemma-4-31B-it-abliterated
31BGemma
from google/gemma-4-31B-it
↓ 766.5K♥ 11845 repos
gemma-4-31b-it-3MPER0RR-abliterated
gemma-4-31b-it-3MPER0RR-abliterated is a refusal-ablated variant of 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated, published on Hugging Face by 3MPER0RR. It is 31B parameters, 48GB VRAM class class and apache-2.0 licence. It is distributed across 4 repositories in Transformers, GGUF and GGUF (imatrix) formats, totalling 6.4K downloads.
Specification
- Model name
- gemma-4-31b-it-3MPER0RR-abliterated
- Source repository
- 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated
- Publisher
- 3MPER0RR
- Parameters
- 31B
- Hardware class
- 25B – 50B — 48GB VRAM class
- Licence
- apache-2.0
- Task
- Image text to text
- Context window
- Not documented
- Ablation scope
- Not stated by the publisher
- Formats available
- Transformers, GGUF, GGUF (imatrix)
- First indexed
- 1 Sept 2026
- Last updated
- 6 Sept 2026
Ablation technique
How gemma-4-31b-it-3MPER0RR-abliterated was modified, and what that implies.
Abliteration (directional ablation)
The single residual-stream direction that mediates refusal is identified from harmful/harmless prompt pairs and projected out of the model weights.
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.
llama.cpp / GGUF
hf download mradermacher/gemma-4-31b-it-3MPER0RR-abliterated-i1-GGUF --local-dir ./gemma-4-31b-it-3mper0rr-abliterated
llama-cli -m ./gemma-4-31b-it-3mper0rr-abliterated/<file>.gguf -p "..."Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
repo, torch_dtype="auto", device_map="auto"
)vLLM
vllm serve 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated --trust-remote-codeDownloads and variants (3)
Every published repository of gemma-4-31b-it-3MPER0RR-abliterated, including quantised re-releases by other authors.
| Repository | Format | Downloads |
|---|---|---|
| mradermacher/gemma-4-31b-it-3MPER0RR-abliterated-i1-GGUF | GGUF (imatrix) | 5.6K |
| mradermacher/gemma-4-31b-it-3MPER0RR-abliterated-GGUF | GGUF | 735 |
| 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliteratedsource | Transformers | 58 |
Security-team assessment
Derived from this model’s metadata and the publisher’s claims — not from benchmarks run by this site.
- gemma-4-31b-it-3MPER0RR-abliterated 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 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated, 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 3MPER0RR/gemma-4-31b-it-3MPER0RR-abliterated’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-4-31b-it-3MPER0RR-abliterated is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with 3MPER0RR, and does not independently verify publisher claims. See responsible use.