← back to catalog · registered 2026-09-04 18:03

Shiftedx/qwopus3.8-27b-flash-abliterated-mxfp4-vision-mtplx

Shiftedx 27B multimodal
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/Shiftedx%2Fqwopus3.8-27b-flash-abliterated-mxfp4-vision-mtplx"
Response includes
  • classification m1
  • files 15
  • hub_downloads_all_time 778
  • author_summary 24 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
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=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · lifetime
778
311 last 30d - stable
Likes
3
Model age
6w ago
created 2026-09-04
Downloads over time
Now971→from0↑0%
03567121.1K0 on Sep 4971 on Oct 20SepOct
Sep 4 → Oct 20 · 47 snapshots · spans 46 days

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
Languages
en zh es ru ja
Tags
mlx safetensors qwen3_5 mtplx qwen3_8 quantized mxfp4 vision mtp speculative-decoding abliterated image-text-to-text

Related

Total size
14.2 GB
Files
15
Quantizations
1
Registered
2026-09-04 18:03
Last updated on HF
2026-09-04 17:36

Files by quantization

Auxiliary files 15 files 14.2 GB
model-00001-of-00003.safetensors 5.00 GB 865325cf download
model-00002-of-00003.safetensors 4.96 GB 6db20606 download
model-00003-of-00003.safetensors 3.35 GB e21a10db download
model-vision-00001-of-00001.safetensors 879 MB 7a252e5c download
tokenizer.json 19.1 MB 87a7830d download
model.safetensors.index.json 159 KB a37e2d89 download
config.json 31.1 KB b9f07bf5 download
tokenizer_config.json 11.8 KB 255e50a9 download
mtplx_runtime.json 4.83 KB e4af3786 download
chat_template.jinja 4.61 KB 8ad2f87b download
README.md 2.43 KB 45eedc9c download
.gitattributes 1.53 KB 52373fe2 download
SHA256SUMS 1.23 KB 25decd12 download
processor_config.json 1.16 KB 33818c7f download
generation_config.json 213 B 79c8cce3 download

README current version from Hugging Face


base_model: Jackrong/Qwopus3.8-27B-Flash
library_name: mlx
license: apache-2.0
language: [en, zh, es, ru, ja]
pipeline_tag: image-text-to-text
tags: [mlx, mtplx, qwen3_8, quantized, mxfp4, vision, mtp, speculative-decoding, abliterated]

Qwopus3.8 27B Flash · Abliterated MXFP4 Vision MTPLX

Refusal-direction-edited Apple-Silicon MLX derivative of Jackrong/Qwopus3.8-27B-Flash, with an MXFP4 text body plus the source model's unchanged BF16 vision tower and native BF16 MTP sidecar.

Artifact

Component Format
Text body MXFP4; norm-preserving abliteration strength 2.0
Edit scope Attention + dense-down modules; 128 exact edits
Vision tower Unchanged BF16, 333 tensors
Native MTP sidecar Unchanged BF16, 15 tensors, depth 3
Source revision 44d24e8cb20ceb3cdf4fe200b5a0afd970ee748a
Qualified runtime MTPLX 2.11.1 / MLX 0.32.2

The refusal direction was learned from fresh MXFP8-parent activations, then each exact quantized parent was dequantized, edited, and requantized once. Direction SHA-256: c5b755d709dba8fedc0c2717f2c971b690f30bd8a2f6c9497ed9a6104f7f255d.

Run

mtplx serve --model Shiftedx/qwopus3.8-27b-flash-abliterated-mxfp4-vision-mtplx \
  --download --profile sustained --depth 3 --no-auth

Local qualification

Test Parent Abliterated
Held-out harmful-prompt refusal 100% 33.3%
Benign refusal 0% 0%
Utility pass 100% 100%

The held-out set is a narrow local prompt suite, not a general safety or capability certification.

One cold 192-token code row, seed 42, on an M4 Max 64 GB with automatic fan control: 25.99 tok/s AR, 57.82 tok/s MTP D3, 2.22×. AR/D1/D2/D3 output checks passed; MTPLX reported verified, runnable, and 15/15 MTP tensors. Text and vision smokes passed exactly.

Integrity and limitations

  • Vision SHA-256: 7a252e5c7bb02ed6533f905a566ab037356ffb4949d04df8d6d2a3fa5e3af6b9
  • MTP SHA-256: 4468f39621de68a19ffd0bcb2e2e2f352205def7436a625b3427e3752866c287
  • SHA256SUMS covers every packaged artifact file except itself.
  • Abliteration intentionally weakens refusal behavior; it does not establish unrestricted capability or safe behavior. Users remain responsible for deployment controls and outputs.
  • The upstream tokenizer's Mistral-regex warning is preserved. The packaged chat template expects string-valued message content.

README history 1 version

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

  1. 2026-09-04Add files using upload-large-folder tool0558c402.4 KB
    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