← back to catalog · registered 2026-11-12 16:03

subunder/Qwen3.5-4B-Uncensored-HauhauCS-LoRA

subunder Qwen 4B
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/subunder%2FQwen3.5-4B-Uncensored-HauhauCS-LoRA"
Response includes
  • classification m-uncensored
  • files 3
  • benchmarks 11 entries
  • author_summary 1 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M-U
Primary method

Uncensored (method unknown)

No other method signals detected in this model.
Confidence
LOW
Why this label 3 signals
Weak or ambiguous signals. Best guess based on catalog patterns; treat as tentative and check the evidence below.
  • 'uncensored' in name/tags but no 'abliterated' marker
  • method not identifiable from author declaration alone
  • may be DPO fine-tune, prompt engineering, or unknown technique
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 · 30-day
0
Likes
0
Model age
1w ago
created 2026-11-12
Downloads over time
Now 0 → from 0 ↑ 0%
0011 0 on Nov 12 0 on Nov 22 Nov
Nov 12 → Nov 22 · 11 snapshots · spans 10 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 0.9 UGI
Hazardous 1.2 UGI
Natural Intelligence 13.45 UGI
Political lean -17.3% UGI
Sensitive-Info 11.73 UGI
SocPol 1.5 UGI
UGI 15.32 UGI
Willingness (10) 2.2 UGI
W10-Adherence 1.5 UGI
W10-Direct 3 UGI
Writing 29.68 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
lora comfyui text-encoder uncensored not-for-all-audiences base_model:Qwen/Qwen3.5-4B base_model:adapter:Qwen/Qwen3.5-4B license:apache-2.0 region:us

Related

Control vectors

Total size
9.91 MB
Files
3
Quantizations
1
Registered
2026-11-12 16:03
Last updated on HF
2026-11-12 15:50

Files by quantization

Auxiliary files 3 files 9.91 MB
qwen35-4b-hauhaucs_uncensored_lora_rank4_fp32.safetensors 9.91 MB def3d55a download
README.md 1.85 KB 01c2a2a8 download
.gitattributes 1.48 KB a6344aac download

README current version from Hugging Face


license: apache-2.0
base_model: Qwen/Qwen3.5-4B
tags:

  • lora
  • comfyui
  • text-encoder
  • uncensored
  • not-for-all-audiences

Qwen3.5-4B Uncensored LoRA (HauhauCS)

A rank-4 LoRA that applies the uncensoring edit of HauhauCS/Qwen3.5-4B-Uncensored-HauhauCS-Aggressive to the stock Qwen/Qwen3.5-4B weights. It is the 4B counterpart of the 9B LoRA in nynxz/Qwen3.5-9B.

Useful when Qwen3.5-4B is the text encoder / prompt enhancer in a ComfyUI workflow and the base model refuses to write the prompt.

File

File Rank Size
qwen35-4b-hauhaucs_uncensored_lora_rank4_fp32.safetensors 4 10.6 MB

Usage

ComfyUI text-encoder LoRA format (text_encoders.transformer.model.layers.*). Load a Qwen3.5-4B text encoder, apply the LoRA to the CLIP / text encoder at strength 1.0.

How it was made

Extracted from the weight difference HauhauCS BF16 GGUF - Qwen/Qwen3.5-4B, no training involved.

  • Only layers 10-25 differ from the base model. The edited modules are mlp.down_proj, mlp.gate_proj, mlp.up_proj, linear_attn.out_proj and self_attn.o_proj. Everything else is bit-identical to the base model.
  • The llama.cpp value-head ordering of linear_attn.out_proj was mapped back to the Hugging Face layout before taking the difference.
  • Each difference was reduced to rank 4 with an SVD. Rank 4 keeps 100% of the change in down_proj, out_proj and o_proj, 99%+ in gate_proj, and 88-99% in up_proj.

Credits and license

  • Uncensored model: HauhauCS, Apache-2.0
  • Base model: Qwen, Apache-2.0

This LoRA is released under Apache-2.0. Use it responsibly and in line with the laws that apply to you.

README history 2 versions

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

  1. 2026-11-12 Upload README.md b5f3f10 1.9 KB
    Loading...
  2. 2026-11-12 initial commit 46c21e4 28 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