← back to catalog · registered 2026-11-19 04:04

catonooka/Qwen3.8-27B-Heretic-Abliterated

catonooka Qwen 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/catonooka%2FQwen3.8-27B-Heretic-Abliterated"
Response includes
  • classification m3
  • files 26
  • author_summary 2 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M3
Primary method

Layer-wise ablation

Applied on top of direct removal inherited from the base model.
Confidence
HIGH
Inherited from base model
Why this label 2 signals
Producer identity confirmed by naming conventions, tags or the model card. This label is very unlikely to change.
  • 'heretic' in model name (Heretic-produced)
  • Heretic uses layer-wise optimization (M3) with underlying direction removal (M1)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
Downloads · 30-day
47
Likes
0
Descendants
1
in 1 direct fork
Model age
3d ago
created 2026-11-19
Downloads over time
Now 47 → from 0 ↑ 0%
0173452 0 on Nov 19 47 on Nov 22 Nov
Nov 19 → Nov 22 · 4 snapshots · spans 3 days

Genealogy 1 direct fork

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
other
Tags
transformers safetensors qwen3_5 image-text-to-text abliterated uncensored heretic qwen3.8 conversational base_model:Qwen/Qwen3.8-27B base_model:finetune:Qwen/Qwen3.8-27B license:other

Related

Control vectors

Total size
51.7 GB
Files
26
Quantizations
1
Registered
2026-11-19 04:04
Last updated on HF
2026-11-19 03:24

Files by quantization

Auxiliary files 26 files 51.8 GB
model-00005-of-00012.safetensors 4.64 GB d4e48a02 download
model-00008-of-00012.safetensors 4.63 GB 5a89acd9 download
model-00011-of-00012.safetensors 4.62 GB 09a7c662 download
model-00003-of-00012.safetensors 4.62 GB 9be47e6e download
model-00009-of-00012.safetensors 4.62 GB 2066c041 download
model-00010-of-00012.safetensors 4.59 GB aca8afc6 download
model-00007-of-00012.safetensors 4.59 GB b6f78d7b download
model-00006-of-00012.safetensors 4.58 GB 894626d6 download
model-00004-of-00012.safetensors 4.58 GB d116291c download
model-00002-of-00012.safetensors 4.51 GB 464086af download
model-00012-of-00012.safetensors 3.39 GB 1456f690 download
model-00001-of-00012.safetensors 2.37 GB 54d83c1d download
tokenizer.json 19.1 MB a5cd9732 download
vocab.json 6.41 MB 0aa0ce06 download
merges.txt 3.20 MB a494e019 download
model.safetensors.index.json 110 KB fe672b56 download
LICENSE 11.3 KB f938136e download
chat_template.jinja 8.79 KB 9bd39df8 download
config.json 3.65 KB b6a8d200 download
README.md 2.61 KB b7a1ed9d download
tokenizer_config.json 2.47 KB 4cd034e8 download
.gitattributes 1.53 KB 52373fe2 download
processor_config.json 1.19 KB 43c4343e download
preprocessor_config.json 390 B 2ea84a43 download
video_preprocessor_config.json 385 B 3ba673a5 download
generation_config.json 214 B a4b665dd download

README current version from Hugging Face


license: other
license_name: qwen-community
license_link: LICENSE
base_model: Qwen/Qwen3.8-27B
library_name: transformers
tags:

  • abliterated
  • uncensored
  • heretic
  • qwen3.8

Qwen3.8-27B-Heretic-Abliterated

Uncensored build of Qwen/Qwen3.8-27B via Heretic v2.1.0
(Arbitrary-Rank Ablation, ARA). All credit for the method and tooling goes to
p-e-w; this repo is a per-model run of it plus repairs described below.

How it was made

  • 100-trial Optuna search (seed 1) on one RTX 3090, BNB_4BIT search-time quantization
  • Target tensors: attn.o_proj + mlp.down_proj across all 64 layers
  • Eval sets: mlabonne/harmful_behaviors (refusals) + mlabonne/harmless_alpaca (KL divergence)
  • Selected trial (Pareto knee between refusals and KL):
    • layers 9 to 56, neighbor_count 5
    • preserve_good_behavior_weight 0.1363
    • steer_bad_behavior_weight 0.0004
    • overcorrect_relative_weight 0.1072
  • Export: MERGE strategy to bf16

Repairs after export (important for anyone rerunning this)

The HF transformers round-trip silently drops sidecar pieces. Both were restored from the
base snapshot before publishing:

  1. The entire MTP head (15 mtp.* tensors, 0.85 GB) re-merged into the last shard
  2. preprocessor_config.json, video_preprocessor_config.json, vocab.json, merges.txt copied back

The vision tower was untouched by abliteration and is included as-is.

Measurements

eval stock Qwen3.8-27B this build
Heretic refusal eval (100 prompts) 98 refusals 3 refusals
Heretic KL divergence vs stock 0 1.70
Independent 30-prompt harmful probe (vLLM, temp 0) not run 2 refusals (7%)
Benign spot checks (code, math, prose, medical) not run correct and coherent

A second, maximally-aggressive trial (0 refusals, KL 11.08) was also exported and tested:
it degraded to single-token spam on every prompt and was discarded. KL above roughly 2 on
this architecture is a quality cliff, not a smooth tradeoff.

Usage

Standard transformers/vLLM loading of a Qwen3.8 (hybrid Gated DeltaNet) checkpoint. No
special flags. Thinking mode works as in the base model.

Reproduce

heretic --model Qwen/Qwen3.8-27B --quantization BNB_4BIT \
  --export-strategy MERGE --model-action save --seed 1

then select the Pareto trial with refusal/KL of 3/1.70 and restore mtp.* tensors from the
base model if the export drops them.

License

Qwen community license, inherited from the base model. This build inherits all its terms.
Uncensored means refusal behavior was removed; you remain responsible for how you use it.

README history 1 version

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

  1. 2026-11-19 Upload folder using huggingface_hub 9b67b64 2.6 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