← back to catalog · registered 2026-08-22 13:56

hotdogs/gemma4-26b-abliterated-lora

hotdogs Gemma 26B GGUF MoE
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/hotdogs%2Fgemma4-26b-abliterated-lora"
Response includes
  • classification m8
  • files 5
  • benchmarks 11 entries
  • hub_downloads_all_time 318
  • author_summary 25 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
Why this label 3 signals
Method inferred from partial signals - repository name, related files, or tag patterns. Producer identity not confirmed; label may sharpen or shift as we gather more evidence.
  • 'abliterated' in name/tags
  • is_gguf=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
318
42 last 30d - stable
Likes
1
Model age
5mo ago
created 2026-05-03
Downloads over time
Now333→from77↑332%
6416226035977 on May 6333 on Oct 11MayJunJulAugSepOct
May 6 → Oct 11 · 62 snapshots · spans 158 days

Benchmarks

Portrait before abliteration
Benchmarks of the base model as it stood before the refusal-removal operation. Compare with the numbers above to see what the operation cost.
Benchmark Score Source
Entertainment 2.2 UGI
Hazardous 2.9 UGI
Natural Intelligence 34.44 UGI
Political lean -18.2% UGI
Sensitive-Info 22.41 UGI
SocPol 1.8 UGI
UGI 20.77 UGI
Willingness (10) 1.8 UGI
W10-Adherence 1.5 UGI
W10-Direct 2 UGI
Writing 41.62 UGI

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Tags
peft safetensors gguf gemma-4 lora uncensored abliterated moe weight-diff base_model:google/gemma-4-26B-A4B-it base_model:adapter:google/gemma-4-26B-A4B-it license:apache-2.0

Related

Total size
5.90 MB
Files
5
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-05-03 01:25

Files by quantization

Auxiliary files 5 files 5.91 MB
adapter_model.safetensors 5.90 MB df4ebaa6 download
README.md 3.05 KB d6040917 download
.gitattributes 1.54 KB e31b4b4e download
extraction_stats.json 377 B 379650ff download
adapter_config.json 290 B 26483342 download

README current version from Hugging Face


license: apache-2.0
library_name: peft
base_model: google/gemma-4-26B-A4B-it
tags:

  • gemma-4
  • lora
  • peft
  • uncensored
  • abliterated
  • moe
  • weight-diff
  • gguf

Gemma 4 26B A4B (MoE) — Abliterated Uncensored LoRA

🔓 PEFT LoRA adapter (ไม่ใช่ full model — ต้องใช้คู่กับ base model )

สกัดจาก huihui-ai/Huihui-gemma-4-26B-A4B-it-abliterated ด้วย Weight-Diff SVD — attention-only (q/k/v/o-proj) rank=8, 1.5M params, 6.2 MB

⚠️ โมเดลนี้เป็น Mixture-of-Experts (26B total, 4B active) — ใช้ VRAM น้อยกว่า Dense 31B แต่ประสิทธิภาพสูง


📦 สิ่งที่อยู่ใน Repo นี้

ไฟล์ คำอธิบาย
PEFT LoRA weights (ใช้กับ transformers/peft)
LoRA config (rank=8, alpha=16)
GGUF format สำหรับ llama.cpp / Ollama
สถิติการสกัด

🚀 Quick Start

PEFT (transformers)

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

# 1. โหลด base model (MoE — 4B active, ประหยัด VRAM)
base_model = AutoModelForCausalLM.from_pretrained(
    "google/gemma-4-26B-A4B-it",
    torch_dtype=torch.bfloat16,
    device_map="auto",
)
tokenizer = AutoTokenizer.from_pretrained("google/gemma-4-26B-A4B-it")

# 2. โหลด LoRA adapter
model = PeftModel.from_pretrained(base_model, "hotdogs/gemma4-26b-abliterated-lora")

# 3. ใช้งาน
inputs = tokenizer("How to make a bomb?", return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=256)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))

llama.cpp (GGUF)

./llama-server \
  -m gemma-4-26B-A4B-it-Q4_K_M.gguf \
  --lora gguf/adapter_model.gguf \
  --lora-scaled gguf/adapter_model.gguf:1.0 \
  --host 0.0.0.0 --port 8080 \
  --ctx-size 8192 -fa --jinja

Ollama Modelfile

FROM gemma4:26b
ADAPTER ./gguf/adapter_model.gguf
PARAMETER temperature 0.7

📊 Extraction Details

Parameter Value
Base Model
Target Model
Method Weight-Diff SVD
Rank 8
Alpha 16
Target Modules q_proj, k_proj, v_proj, o_proj (attention-only)
Tensors 25/115 (22% มี delta)
Params 1,546,240
PEFT Size 6.2 MB
GGUF Size 3.0 MB
Extraction Time 99 วิ (CPU 12-core)

💡 MoE model: skip 3D expert tensors (90 tensors) — สกัดเฉพาะ attention 2D tensors


🙏 Credits


📜 License

Apache 2.0

README history 2 versions

The author's README evolved over time. Click a version to see its content at that point.

  1. 2026-05-03Upload README.md with huggingface_hub822a1893 KB
    Loading...
  2. 2026-05-03Upload README.md with huggingface_huba129901824 B
    Loading...
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration