CUSTOM
CLANKERS.
← Browse all models

DavidAU / Model details

Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source

DavidAU/Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source

abliterateduncensoredfine-tuned
Download files on Hugging Face ↗

Read the original model card ↗ · Choosing a model

Summarize and ask questions with AI

About this model

The repository Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source was created by DavidAU and is derived from trohrbaugh/Qwen3.8-27B-heretic-ara. According to the model card the creator says the model was abliterated and uncensored, further trained on a horror dataset and equipped with a custom "Halloween O/S" that provides a set of ten evil AI constructs and 13+13 instruct and reasoning modes for role‑play. The weights are provided in safetensors format and the model is released under the Apache‑2.0 license. The creator describes the model as an “evil all the time” horror‑themed system that can summon characters such as Joker, Freddy Krueger, or any user‑specified persona, and notes that it supports image‑text input via an accompanying mmproj file. The model card also warns that the system is intended for entertainment only, suggests a minimum context of 64 k tokens for long conversations, and advises against enabling tool use or taking the generated content at face value.

Description sources

Source-based description, not an independent evaluation. Check the linked sources for the creator’s claims and limitations.

Size and formats

Base model
trohrbaugh/Qwen3.8-27B-heretic-ara
Parameters
27.3567B
Main repository formats
safetensors
License tag
apache-2.0
Languages
en
Last updated
2026-10-30 · 05:46:29.000 UTC

What changed

Model types are creator-reported, not independently tested. These labels describe source metadata, not a guarantee of behavior.

  • abliteratedsource tag: abliterated
  • uncensoredsource tag: uncensored
  • uncensoredmodel name: Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source
  • fine-tunedsource tag: base_model:finetune:trohrbaugh/Qwen3.8-27B-heretic-ara

Family evidence

These groupings are source-reported. Explicit base identifiers take priority over generic architecture tags.

  • Qwen: Base model: trohrbaugh/Qwen3.8-27B-heretic-ara

Available versions

Versions appear here only when they explicitly name this exact model as their source. Check each repository for available files.

Refusal profile

Not yet measured: no GGUF version is available for our runner.

How we measure

Link to this listing

Made or quantized this model? Paste one of these into your model card or README so people can find this listing.

Markdown

[Listed on Custom Clankers](https://customclankers.com/models/davidau-qwen3-8-27b-halloween-terror-heretic-uncensored-source/)

HTML

<a href="https://customclankers.com/models/davidau-qwen3-8-27b-halloween-terror-heretic-uncensored-source/">Listed on Custom Clankers</a>

Plain text links only: no image, no script, nothing loaded from this site.

Customize or build on this model

Listed from public GitHub data, not tested by us. Read the code and license before you run anything.

Tools to customize it further

Remove more refusals

  • Heretic ↗ ★ 37,081

    Removes refusal behavior from transformer models automatically. It pairs directional ablation (abliteration) with an Optuna parameter search that tries to cut refusals while keeping the model close to the original, so you don't have to pick layers by hand.

  • OBLITERATUS ↗ ★ 9,021

    A larger refusal-removal toolkit with several abliteration methods, usable locally or through a hosted Hugging Face Space. Check the telemetry setting before you run it: the README describes telemetry as opt-in, but on by default in the Spaces version.

Fine-tune it on your own data

  • Unsloth ↗ ★ 78,752

    A desktop app and Python library for running and training models locally. Its README lists LoRA, QLoRA, full fine-tuning, DPO and GRPO, plus export to formats including GGUF, on Windows, Linux, WSL and macOS.

  • LlamaFactory ↗ ★ 75,778

    Fine-tunes more than 100 language and vision-language models from a command line or a browser-based Web UI, aimed at training without writing code.

Make a smaller download (quantize)

  • llama.cpp ↗ ★ 132,761

    Runs models locally and includes the tools behind most GGUF files. A conversion script turns Hugging Face weights into GGUF, and llama-quantize shrinks a high-precision GGUF to lower-bit formats, which can cost some accuracy.

  • MLX LM ↗ ★ 7,503

    Apple's MLX package for running and fine-tuning models on Apple silicon. It can quantize models and upload them to the Hugging Face Hub, and supports LoRA and full fine-tuning.

Merge it with another model

  • mergekit ↗ ★ 7,421

    Merges pretrained models with many merge methods. It works out-of-core, so merges can run on a CPU or with as little as 8 GB of VRAM, and it also handles LoRA extraction and mixture-of-experts merges.

Learn the techniques

  • LLM Course ↗ ★ 84,760

    A free course with roadmaps and Colab notebooks. Its notebooks include one-click tools for merging with MergeKit, quantizing to GGUF, GPTQ, EXL2, AWQ and HQQ, and automatic abliteration with custom datasets.

Compare all tools →

Source provenance

Downloads
0
Likes
0
Metadata checked
2026-10-30 · 08:12:02 UTC
First seen in index
2026-10-30 · 08:12:02 UTC
Source revision
8ec84ae7e8539343a18318a6a7c030116314aa05

Source-card excerpt

Source model card excerpt: Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source Source code for GGUFS (NEO/FULL MTP): --- Constructed with the "Halloween O/S" (low, medium and xhigh) for your "Trick or Treat" pleasure featuring

From DavidAU/Qwen3.8-27B-Halloween-Terror-Heretic-Uncensored-source. Creator model-card excerpt, not an independent review.

Inspect before you run

Low refusal is not proof of correctness. Outputs can be inaccurate, harmful, or biased. Review the model card, license and files yourself. A license tag is not commercial-use advice. Parameter counts are source metadata; no hardware compatibility is implied.

This directory does not host or redistribute model weights. Refusal profiles are measured by us on a fixed prompt set; all other behavior claims are source-reported.

Collection rules & limitations ↗