⌦abliteratedmodels.org

Gemma-4-26B-A4B-Plainspeak-SOMPOA

Gemma-4-26B-A4B-Plainspeak-SOMPOA is a Heretic-ablated variant of 26B-Suite/BROKEN_Gemma-4-26B-A4B-Plainspeak-SOMPOA, published on Hugging Face by mradermacher. It is 26B total parameters (mixture-of-experts), 48GB VRAM class class and apache-2.0 licence. It is distributed across 3 repositories in GGUF and GGUF (imatrix) formats, totalling 10.6K downloads.

Heretic26B MoEGemma25B – 50B↓ 10,639
View on Hugging Face ↗Base model ↗

Specification

Model name
Gemma-4-26B-A4B-Plainspeak-SOMPOA
Source repository
mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-i1-GGUF
Base model
26B-Suite/BROKEN_Gemma-4-26B-A4B-Plainspeak-SOMPOA
Publisher
mradermacher
Parameters
26B (mixture-of-experts)
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
GGUF, GGUF (imatrix)
First indexed
16 Aug 2026
Last updated
18 Aug 2026

Ablation technique

How Gemma-4-26B-A4B-Plainspeak-SOMPOA 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

“For a convenient overview and download list, visit our model page for this model.”

Running it

Commands are templates — confirm the exact repository and quant file before use.

llama.cpp / GGUF
hf download mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-i1-GGUF --local-dir ./gemma-4-26b-a4b-plainspeak-sompoa
llama-cli -m ./gemma-4-26b-a4b-plainspeak-sompoa/<file>.gguf -p "..."
Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

repo = "mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-i1-GGUF"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(
    repo, torch_dtype="auto", device_map="auto"
)
vLLM
vllm serve mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-i1-GGUF --trust-remote-code

Downloads and variants (2)

Every published repository of Gemma-4-26B-A4B-Plainspeak-SOMPOA, including quantised re-releases by other authors.

RepositoryFormatDownloads
mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-i1-GGUFsourceGGUF (imatrix)9.4K
mradermacher/Gemma-4-26B-A4B-Plainspeak-SOMPOA-GGUFGGUF1.3K

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-Plainspeak-SOMPOA 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 26B-Suite/BROKEN_Gemma-4-26B-A4B-Plainspeak-SOMPOA, 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 26B-Suite/BROKEN_Gemma-4-26B-A4B-Plainspeak-SOMPOA’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-Plainspeak-SOMPOA is collected from the public Hugging Face API and the publisher’s model card. This site does not host weights, is not affiliated with mradermacher, and does not independently verify publisher claims. See responsible use.