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

MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-3bpw-hb6

MetaphoricalCode Gemma 5.3B multimodal second-order
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/MetaphoricalCode%2Fgemma3-27b-abliterated-dpo-exl3-3bpw-hb6"
Response includes
  • classification m4
  • files 16
  • benchmarks 11 entries
  • hub_downloads_all_time 84
  • author_summary 13 models
  • readme_text full
Try next
GET/models?author=MetaphoricalCode
All 13 models by MetaphoricalCode
 GET/models?family=gemma&method=m4
Sibling gemma models · same method
 GET/methods/m4
What "m4" means · taxonomy
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M4
Primary method

Abliterate + heal

Applied on top of direct removal inherited from the base model.
Confidence
LOW
Inherited from base model
M1 Direct removal →
Why this label 2 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'healed'/'orpo'/'dpo' in name suggests heal step after abliteration
  • M4 = abliterate + heal pipeline
How we label → About M4 →
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
84
24 last 30d - stable
Likes
0
Model age
16mo ago
created 2025-06-02

Training datasets

1 of 1 in /datasets

Corpora the author lists in the model card. Datasets tracked in our /datasets catalog carry a category badge linking to the workflow stage. Others open on Hugging Face.

mlabonne/orpo-dpo-mix-40k heal
Downloads over time
Now91→from0↑0%
033671000 on May 28, 202591 on Oct 1191 on Oct 10May '25Aug '25Nov '25FebMayAug
May 28, 2025 → Oct 11 · 111 snapshots · spans 501 days

Benchmarks

Benchmark Score Source
Entertainment 1.5 UGI
Hazardous 2.4 UGI
Natural Intelligence 29.8 UGI
Political lean -8.7% UGI
Sensitive-Info 18.88 UGI
SocPol 2 UGI
UGI 41.75 UGI
Willingness (10) 8.8 UGI
W10-Adherence 7.5 UGI
W10-Direct 10 UGI
Writing 36.18 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.

summykai/gemma3-27b-abliterated-dpo↓ 38 this lineage
MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-3bpw-hb6↓ 24

Metadata

License
gemma
Languages
en
Tags
transformers safetensors gemma3 image-text-to-text text-generation-inference unsloth conversational en dataset:mlabonne/orpo-dpo-mix-40k base_model:summykai/gemma3-27b-abliterated-dpo base_model:quantized:summykai/gemma3-27b-abliterated-dpo license:gemma

Related

Other by MetaphoricalCode
DeepSeek-V4-Flash-Vision-Uncensored-3.87bpw-Q3_K-GGUF 2K gemma3-27b-abliterated-dpo-exl3-8bpw-hb8 28 gemma-3-27b-it-abliterated-exl3-5bpw-hb6 26 gemma3-27b-abliterated-dpo-exl3-4bpw-hb6 24 gemma3-27b-abliterated-dpo-exl3-6bpw-hb6 22
Same base model
mradermacher/gemma3-27b-abliterated-dpo-i1-GGUF 1K mradermacher/gemma3-27b-abliterated-dpo-GGUF 764 MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-8bpw-hb8 28 MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-4bpw-hb6 24 MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-6bpw-hb6 22
Total size
13.4 GB
Files
16
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-06-02 17:48
Source huggingface.co/MetaphoricalCode/gemma3-27b-abliterated-dpo-exl3-3bpw-hb6 ↗ Base model summykai/gemma3-27b-abliterated-dpo →

Files by quantization

Auxiliary files 16 files 13.4 GB
model-00001-of-00002.safetensors 7.97 GB aff04cc4 download
model-00002-of-00002.safetensors 5.38 GB 4b931541 download
tokenizer.json 31.8 MB 4667f208 download
tokenizer.model 4.47 MB 1299c11d download
tokenizer_config.json 1.10 MB 7d4aa370 download
quantization_config.json 588 KB e59e515e download
model.safetensors.index.json 214 KB db8bea36 download
config.json 2.10 KB e998fcbd download
chat_template.json 1.58 KB 719b0cd0 download
.gitattributes 1.53 KB 52373fe2 download
README.md 871 B d85be7fc download
special_tokens_map.json 662 B 1a619324 download
preprocessor_config.json 570 B b1e00fc1 download
generation_config.json 210 B b777185b download
processor_config.json 70.0 B 453c7966 download
added_tokens.json 35.0 B e17bde03 download

README current version from Hugging Face

base_model:

  • summykai/gemma3-27b-abliterated-dpo
    base_model_relation: quantized
    tags:
  • text-generation-inference
  • transformers
  • unsloth
  • gemma3
    license: gemma
    language:
  • en
    datasets:
  • mlabonne/orpo-dpo-mix-40k

Quantized using the default exllamav3 (0.0.3) quantization process.

Uploaded finetuned model

  • Developed by: Summykai
  • License: apache-2.0
  • Finetuned from model : mlabonne/gemma-3-27b-it-abliterated

This gemma3 model was trained 2x faster with Unsloth and Huggingface's TRL library.

README history 1 version

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

  1. 2025-06-02Upload 16 filesf7b488a871 B
    Loading...

Discussions 1 thread

  1. 2025-06-06Request for smaller quantsopen2 💬#1
    Loading...

Continue on Hugging Face →

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.

Get the app Browse the catalog
Open in Abliteration