Get the app
← back to catalog · registered 2026-11-07 14:02

moviestation/Huihui-Qwen3.5-4B-abliterated-MNN

moviestation 4B 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/moviestation%2FHuihui-Qwen3.5-4B-abliterated-MNN"
Response includes
  • classification m1
  • files 11
  • author_summary 2 models
  • readme_text full
Try next
GET/models?author=moviestation
All 2 models by moviestation
 GET/methods/m1
What "m1" means · taxonomy
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)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-11-07

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.

huihui-ai/Huihui-Qwen3.5-4B-abliterated↓ 19,849 this lineage
moviestation/Huihui-Qwen3.5-4B-abliterated-MNN

Metadata

License
apache-2.0
Tags
mnn quantized android image-text-to-text base_model:huihui-ai/Huihui-Qwen3.5-4B-abliterated base_model:finetune:huihui-ai/Huihui-Qwen3.5-4B-abliterated license:apache-2.0 region:us

Related

Other by moviestation
Huihui-Qwen3.5-2B-abliterated-MNN 0
Same base model
mradermacher/Huihui-Qwen3.5-4B-abliterated-GGUF 89K mradermacher/Huihui-Qwen3.5-4B-abliterated-i1-GGUF 4K interimlabs/InterimLabs-Huihui-Qwen3.5-4B-abliterated-GGUF-Q4_K_M 3K darkmaniac7/Qwen3.5-4B-uncensored-MNN 1K interimlabs/InterimLabs-Huihui-Qwen3.5-4B-abliterated-MLX-4bit 681
Total size
0 B
Files
11
Quantizations
1
Registered
2026-11-07 14:02
Last updated on HF
2026-11-07 14:02
Source huggingface.co/moviestation/Huihui-Qwen3.5-4B-abliterated-MNN ↗ Base model huihui-ai/Huihui-Qwen3.5-4B-abliterated →

Files by quantization

Auxiliary files 11 files 2.55 GB
llm.mnn.weight 2.20 GB 6d74453d download
visual.mnn.weight 346 MB c81ce3d9 download
tokenizer.mtok 6.99 MB 37ff2b03 download
llm.mnn 3.85 MB 5cafb34c download
visual.mnn 712 KB 008abfa3 download
LICENSE 11.3 KB f938136e download
llm_config.json 9.22 KB 3be79372 download
.gitattributes 1.72 KB f081f536 download
release-manifest.json 1.27 KB b6d95f86 download
README.md 946 B 2ebab64c download
config.json 661 B 69ad85e8 download

README current version from Hugging Face

license: apache-2.0
base_model: huihui-ai/Huihui-Qwen3.5-4B-abliterated
pipeline_tag: image-text-to-text
tags:

  • mnn
  • quantized
  • android

Huihui-Qwen3.5-4B-abliterated — MNN conversion

Converted for the RH Studio Android local assistant. Original model by
huihui-ai, based on Qwen3.5-4B.
This repository contains a format conversion, not an independently trained model.

Source revision: 5581467dfd52bf338c782006a6cdce05c42594be.
Changes: MNN export, 4-bit text weights (block size 64), 8-bit visual weights.
Original Apache 2.0 license is included in LICENSE.

Download separately from the app. These weights must not be bundled into the APK.
See release-manifest.json for file sizes and SHA-256 hashes.
Android performance and memory use depend on the device; conversion alone does
not establish on-device compatibility or output quality.

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