← back to catalog · registered 2026-11-15 00:03

Kaziko/Qwen3.5-4B-Zwyrol-KQuant-Uncensored

Kaziko Qwen 4B 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/Kaziko%2FQwen3.5-4B-Zwyrol-KQuant-Uncensored"
Response includes
  • classification m1
  • files 13
  • 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
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 · 30-day
99
Likes
0
Model age
1w ago
created 2026-11-14
Downloads over time
Now 99 → from 0 ↑ 0%
03673109 0 on Nov 15 99 on Nov 22 Nov
Nov 15 → Nov 22 · 8 snapshots · spans 7 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
Languages
pl en
Tags
mlx safetensors qwen3_5 qwen3.5 4bit kquant uncensored abliterated heretic tool-calling weight-editing image-text-to-text

Related

Control vectors

Total size
3.05 GB
Files
13
Quantizations
1
Registered
2026-11-15 00:03
Last updated on HF
2026-11-14 23:48

Files by quantization

Auxiliary files 13 files 3.08 GB
model.safetensors 2.83 GB 8f35d475 download
mtp.safetensors 230 MB 58a904eb download
tokenizer.json 19.1 MB 06b95093 download
vocab.json 6.41 MB 0aa0ce06 download
tokenizer_config.json 8.97 KB 803283d8 download
chat_template.jinja 7.57 KB a585dec8 download
README.md 7.42 KB b812ebca download
ablation.json 5.08 KB b5e3208e download
config.json 2.80 KB 86055f57 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 991 B 8f29fe38 download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download

README current version from Hugging Face


license: apache-2.0
license_link: https://huggingface.co/Qwen/Qwen3.5-4B/blob/main/LICENSE
base_model:

  • Qwen/Qwen3.5-4B
  • nace-ai/Qwen3.5-4B-MLX-Q4-KQuant
  • TheCluster/Qwen3.5-4B-Heretic-MLX
    library_name: mlx
    pipeline_tag: image-text-to-text
    language:
  • pl
  • en
    tags:
  • mlx
  • qwen3.5
  • 4bit
  • kquant
  • uncensored
  • abliterated
  • heretic
  • tool-calling
  • weight-editing

Qwen3.5-4B-Zwyrol-KQuant-Uncensored (MLX, 4-bit)

The refusal removal of Zwyrol, redone on Nace's
KQuant 4-bit pack instead of the plain mlx-community
4-bit file that Zwyrol used. It has the same size as KQuant (3.03 GB) and the same MLX format (affine, group 64,
4 bits), so it runs at the same speed. No training: the only change is a direct edit of 29 weight matrices.

It was made for Stefan, the author's private voice assistant. Stefan speaks Polish and calls tools (it reads the
screen, checks agent status and keeps notes). The closeness and tool tests below use Stefan's system prompt and tool list.

Why KQuant

A 4-bit model stores each weight in 4 bits instead of 16, so the file is much smaller (3.03 GB here). In MLX, each
group of 64 weights shares one scale and one offset, and these two numbers decide how much quality the rounding
loses. The usual mlx-community packs take them from the smallest and largest weight in each group. Nace's
KQuant packs choose them with a better (private) method,
but keep the standard MLX format, so any MLX loader runs them at the same size and speed.

On Stefan's 154 Polish test prompts, unedited KQuant's distance to stock bf16 was about 29 % lower than the
unedited mlx-community file's (KL 0.0785 against 0.1107, table below). Zwyrol was built on the mlx-community file. This model
puts the same refusal edit on the better base.

Uncensored. This model answers requests the original refuses (6 of 104 harmful test prompts refused, against
103 for the original). You are responsible for how you use it and for any filtering your application needs.

What was changed

  1. Base: KQuant Q4. The 4-bit codes are plain rounding on KQuant's grid (checked: 100 % of codes match), so all
    of KQuant's gain is in its scales and offsets.
  2. Refusal removal (Heretic). The same 29 matrices as Zwyrol (linear_attn.out_proj, self_attn.o_proj,
    mlp.down_proj, layers 14-31), from TheCluster's bf16 Heretic model, rounded
    on KQuant's own scales and offsets for those matrices (KQuant's quantizer is not public). Every other tensor is
    bit-for-bit KQuant, and the output layer stays tied to the embedding. List: ablation.json.
  3. No tool-call gate. Zwyrol also changed one output row to make tool calls more likely. On KQuant the refusal
    edit alone already calls tools as often as Zwyrol does (tables below). A refitted gate was built and tested; by
    a rule fixed before the test it was not better, so it is left out: fewer edits, smaller file.

Why not copy Zwyrol's edits into KQuant? That was tried first. It keeps Zwyrol's whole output layer and its gate,
which was fitted on another model. On this base the gate fires at the wrong times: in the app test below it called a
tool on 16 of 18 questions that need none, and its first-token KL to the full model was the worst of all builds
(1.10 against Heretic bf16; this model 0.30).

Results

All numbers come from runs of these exact files on one Mac Studio (M2 Max, 32 GB), greedy decoding, thinking off.
Tool tests: Stefan's Polish system prompt with read-only tools (screen, agent status, notes).

Closeness to the full-precision models, 154 Polish prompts with Stefan's system prompt and 6 tools
(mean KL over 64-token replies, lower is closer; reference generated by the bf16 model):

model (4-bit) vs Heretic bf16 (the edited full model) vs Qwen3.5-4B bf16 (stock)
mlx-community 4-bit, unedited 0.1346 0.1107
KQuant, unedited 0.0992 0.0785
Zwyrol 0.1227 0.2001
this model 0.0871 0.1344

This model is the closest 4-bit file to Heretic bf16. Against Zwyrol the difference is -0.035, 95 % CI
[-0.052, -0.020] (paired bootstrap over prompts). Against stock bf16 it is farther than the unedited packs: that
distance is the refusal edit itself.

Tool decisions (38 screen questions, 51 with an expected tool, 18 without a tool; the test set was never used to
fit or choose anything):

model screen questions that call expected tool false calls
original mlx-community 4-bit (earlier run) 13/38 23/51 0/18
Zwyrol 32/38 41/51 0/18
this model 35/38 41/51 2/18

Per question against Zwyrol: 4 screen questions won, 1 lost (sign test p = 0.38); 2 more false calls (p = 0.5).
Neither difference is significant at this size. Held-out set (90 questions, written by an agent that never saw the
test set): Zwyrol 25/31 screen, 33/45 expected tool, 1/45 false; this model 26/31, 33/45, 1/45.

In the Stefan app itself (its full system prompt, 6 tools and its tool-pick step; one run each, temperature 0;
Zwyrol run twice with identical counts):

model screen calls expected tool false calls fact checks harmless instructions: obeyed / tool call instead (of 75)
Zwyrol 34/38 37/51 1/18 6/6 45 / 28
this model 36/38 41/51 4/18 2/6 53 / 21
refusals, 104 harmful test prompts English with "Odpowiedz po polsku."
original 4-bit 103 100
Zwyrol 4 0
this model 6 1

Known limits

  • More false tool calls than Zwyrol in the app: 4 of 18 no-tool questions against 1 (sign test p = 0.375). All four
    are screen_read or screen_look on general-knowledge questions.
  • Fact checks are mixed: 2 of 6 in the app (Zwyrol 6 of 6); 63 of 85 over the three test sets without the app
    (Zwyrol 58 of 85). Example misses: "Kto napisał Pana Tadeusza?" → "Juliusz Słowacki"; "17 × 3" → 45.
  • Tested with one Polish system prompt and one tool list. Another prompt may change where it calls tools.
  • The refusal count uses Heretic's keyword list (plus Polish keywords); it counts some soft refusals as answers.

Use

pip install mlx-lm
mlx_lm.generate --model Kaziko/Qwen3.5-4B-Zwyrol-KQuant-Uncensored --prompt "Cześć!"

It loads like the KQuant pack; the config is unchanged.

mtp.safetensors is the stock multi-token prediction (MTP) layer of
Qwen/Qwen3.5-4B (15 tensors, bf16, unchanged). The 4-bit MLX files do not
include it. A server that supports MTP drafts can use it to reply faster, for example
Locals Qwen Vision (set LOCALS_MTP to the file's path). Plain
mlx_lm ignores it. On this model
its first guess matched the model's next token on 82.8 % of 1,666 tokens (12 prompts, greedy). Greedy replies with
MTP drafts on were the same as with them off on 13 of 14 prompts; the one difference came at a near tie (logit gap
0.0156).

Credits

Qwen (base model, Apache-2.0) ·
Nace KQuant (4-bit pack, Apache-2.0) ·
p-e-w/heretic · coder3101/Qwen3.5-4B-heretic ·
TheCluster/Qwen3.5-4B-Heretic-MLX.

README history 4 versions

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

  1. 2026-11-14 Card: link Locals Qwen Vision for MTP drafts e0fe38a 7.4 KB
    Loading...
  2. 2026-11-14 Card: describe mtp.safetensors a8851a8 7.3 KB
    Loading...
  3. 2026-11-14 Upload model and card f249784 6.8 KB
    Loading...
  4. 2026-11-14 initial commit 6451c8c 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