← back to catalog · registered 2026-11-15 14:03

paloik/Qwen3.8-27B-Uncensored-Heretic-MTP-int4-ov

paloik 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/paloik%2FQwen3.8-27B-Uncensored-Heretic-MTP-int4-ov"
Response includes
  • classification m3
  • files 2
  • 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
0
Likes
0
Model age
6d ago
created 2026-11-15
Downloads over time
Now 0 → from 0 ↑ 0%
0011 0 on Nov 15 0 on Nov 22 Nov
Nov 15 → Nov 22 · 8 snapshots · spans 7 days

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
openvino int4 mtp multi-token-prediction abliterated uncensored heretic qwen3.8 coding image-text-to-text base_model:Qwen/Qwen3.8-27B base_model:finetune:Qwen/Qwen3.8-27B

Related

Control vectors

Total size
0 B
Files
2
Quantizations
1
Registered
2026-11-15 14:03
Last updated on HF
2026-11-15 13:24

Files by quantization

Auxiliary files 2 files 3.04 KB
README.md 1.56 KB f88c3e9a download
.gitattributes 1.48 KB a6344aac download

README current version from Hugging Face


license: apache-2.0
base_model:

  • Qwen/Qwen3.8-27B
  • llmfan46/Qwen3.8-27B-Uncensored-Heretic-Native-MTP-Preserved
    tags:
  • openvino
  • int4
  • mtp
  • multi-token-prediction
  • abliterated
  • uncensored
  • heretic
  • qwen3.8
  • coding
    library_name: openvino
    pipeline_tag: image-text-to-text

Qwen3.8-27B-Uncensored-Heretic-MTP (OpenVINO INT4)

OpenVINO IR (INT4) quantized version of llmfan46/Qwen3.8-27B-Uncensored-Heretic-Native-MTP-Preserved, optimized for Intel Core Ultra / Arc GPUs using OpenVINO GenAI (2026.4+) with native Multi-Token Prediction (MTP) preserved.

Quickstart (OpenVINO GenAI with MTP)

import openvino_genai as ov_genai

model_dir = "./qwen3.8-27b-uncensored-mtp-int4-ov"
draft_model = ov_genai.draft_model(model_dir, "GPU")

pipe = ov_genai.VLMPipeline(
    model_dir,
    "GPU",
    draft_model=draft_model,
    scheduler_config=ov_genai.SchedulerConfig(enable_prefix_caching=False),
    CACHE_DIR="./ov_cache"
)

config = ov_genai.GenerationConfig()
config.max_new_tokens = 1024
config.num_assistant_tokens = 5
config.temperature = 0.0

pipe.start_chat("You are an intelligent, helpful, and direct AI assistant. Answer all user requests objectively, concisely, and without unnecessary lecturing.")
print(pipe.generate("Write a high-performance async file watcher in Python.", config))
pipe.finish_chat()

License & Disclaimer

Licensed under Apache 2.0. Abliterated research model with reduced refusals; users are responsible for verifying outputs and complying with applicable laws.

README history 1 version

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

  1. 2026-11-15 Upload README.md with huggingface_hub 028b289 1.6 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