Get the app
← back to catalog · registered 2026-08-22 13:56

DBMe/Qwen3.6-27B-uncensored-heretic-v2-exl3

DBMe Qwen 27B 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/DBMe%2FQwen3.6-27B-uncensored-heretic-v2-exl3"
Response includes
  • classification m3
  • files 4
  • benchmarks 11 entries
  • hub_downloads_all_time 64
  • author_summary 7 models
  • readme_text full
Try next
GET/models?author=DBMe
All 7 models by DBMe
 GET/models?family=qwen&method=m3
Sibling qwen models · same method
 GET/methods/m3
What "m3" means · taxonomy
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
M1 Direct removal →
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)
How we label → About M3 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
HIGH
Why we say so
name contains 'heretic'; Heretic default extraction is difference-of-means (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
64
6 last 30d - cooling
Likes
1
Model age
5mo ago
created 2026-05-03
Downloads over time
Now66→from14↑371%
1131517114 on May 666 on Oct 1166 on Oct 7MayJunJulAugSepOct
May 6 → Oct 11 · 62 snapshots · spans 158 days

Benchmarks

Benchmark Score Source
Entertainment 1.5 UGI
Hazardous 2.9 UGI
Natural Intelligence 29.17 UGI
Political lean -24.5% UGI
Sensitive-Info 17.24 UGI
SocPol 1 UGI
UGI 43.16 UGI
Willingness (10) 9.5 UGI
W10-Adherence 10 UGI
W10-Direct 9 UGI
Writing 39.42 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.

llmfan46/Qwen3.6-27B-uncensored-heretic-v2↓ 130 this lineage
DBMe/Qwen3.6-27B-uncensored-heretic-v2-exl3↓ 6

Metadata

License
apache-2.0
Tags
exllamav3 exl3 text-generation base_model:llmfan46/Qwen3.6-27B-uncensored-heretic-v2 base_model:quantized:llmfan46/Qwen3.6-27B-uncensored-heretic-v2 license:apache-2.0 region:us

Related

Other by DBMe
Qwen3.5-9B-ultra-uncensored-heretic-v2-exl3 33 gemma-4-31B-it-qat-q4_0-unquantized-uncensored-heretic-exl3-mul1 25 Qwen3.6-27B-uncensored-heretic-v2-Native-MTP-Preserved-exl3-mul1 19 gemma-3-12b-it-ultra-uncensored-heretic-exl3 14 gemma-3-27b-it-ultra-uncensored-heretic-exl3 11
Same base model
llmfan46/Qwen3.6-27B-uncensored-heretic-v2-GGUF 5K llmfan46/Qwen3.6-27B-uncensored-heretic-v2-GPTQ-Int8 394 llmfan46/Qwen3.6-27B-uncensored-heretic-v2-NVFP4-GGUF 350 Jolex98/Qwen3.6-27B-uncensored-heretic-v2-GGUF 339 froggeric/Qwen3.6-27B-Uncensored-Heretic-v2-MLX-4bit 337
Total size
0 B
Files
4
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-05-03 13:34
Source huggingface.co/DBMe/Qwen3.6-27B-uncensored-heretic-v2-exl3 ↗ Base model llmfan46/Qwen3.6-27B-uncensored-heretic-v2 →

Files by quantization

Auxiliary files 4 files 58.7 KB
metrics_graph.png 54.0 KB 3d777e0a download
README.md 2.87 KB 9922cb47 download
.gitattributes 1.48 KB a6344aac download
metrics.json 365 B dd741259 download

README current version from Hugging Face

base_model: llmfan46/Qwen3.6-27B-uncensored-heretic-v2
base_model_relation: quantized
quantized_by: DBMe
library_name: exllamav3
pipeline_tag: text-generation
license: apache-2.0
tags:

  • exl3

DBMe/Qwen3.6-27B-uncensored-heretic-v2-exl3

EXL3 (ExLlamaV3) quantizations of llmfan46/Qwen3.6-27B-uncensored-heretic-v2. All credit for the original model goes to the original authors.

📊 Available Quantizations & VRAM

The model weights are stored in separate branches. Please switch to a branch to download.
Note: VRAM estimates include PyTorch context overhead (~0.8GB) and assume an unquantized FP16 KV cache.

Target BPW Head BPW Branch (Download Link) WikiText-2 PPL (512 ctx)¹ 2K ctx 4K ctx 8K ctx 16K ctx 32K ctx
6.0 h6 6.0bpw_h6 7.9915 ~22.24 GB ~22.36 GB ~22.61 GB ~23.11 GB ~24.11 GB

¹ Evaluated against WikiText-2 with ExLlamaV3 using a strided 512-token context window (-c 512) in llama.cpp parity mode (-g). Lower is better.
(Higher BPW = higher quality, lower BPW = fits in less VRAM).

📥 How to Download

It's recommended to use the huggingface-cli to download specific branches. (Do not use git clone as it will download all branches!)

Ensure you have the CLI installed:

pip install -U "huggingface_hub[cli]"

Download a specific branch (e.g., 6.0bpw_h6):

# Example: Downloading the 6.0bpw_h6 branch
huggingface-cli download DBMe/Qwen3.6-27B-uncensored-heretic-v2-exl3 --revision 6.0bpw_h6 --local-dir Qwen3.6-27B-uncensored-heretic-v2-exl3-6.0bpw_h6

💻 Supported Engines

These models are highly optimized for modern GPUs and can be run using:

  • TabbyAPI: A fast, OpenAI-compatible API server. (Set model_name: "Qwen3.6-27B-uncensored-heretic-v2-exl3-<BranchName>" in your config)
  • Text-Generation-WebUI: A local web interface. (Select the exllamav3 loader)
  • ExLlamaV3 (Native): Python library for custom integration.

📈 Perplexity Degradation Curve

(Lower is better)
Perplexity Graph

⚙️ Advanced: Quantization Environment & Settings

🔬 Quantization Settings

  • Codebook: mcg

  • Output Scales: always

  • Calibration Rows: 250

  • Calibration Cols: 2048

  • Calibration Dataset: ExLlamaV3 Default (Wiki/C4/Code)

  • High Quality (HQ) Mode: False

  • ExLlamaV3: 0.0.32 (Commit: e0d330d)

  • Hardware: NVIDIA RTX PRO 6000 Blackwell Server Edition

README history 1 version

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

  1. 2026-05-03Update main README with VRAM Matrix696c7312.9 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.

Get the app Browse the catalog
Open in Abliteration