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

hotdogs/qwen35b-a3b-fable-sft-abliterated

hotdogs Qwen GGUF MoE second-order 262K ctx
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/hotdogs%2Fqwen35b-a3b-fable-sft-abliterated"
Response includes
  • classification m8
  • files 3
  • hub_downloads_all_time 9,471
  • author_summary 25 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M8
Primary method

Repackaging (quantization)

Applied on top of direct removal inherited from the base model.
Confidence
MEDIUM
Inherited from base model
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=1
  • assume M1 (base ablation) + M8 (GGUF quant) - default when producer unknown
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · lifetime
9K
367 last 30d - cooling
Likes
15
Model age
3mo ago
created 2026-06-30

Training datasets

Corpora the author lists in the model card. Datasets tracked in our /datasets catalog carry a category badge linking to the workflow stage. Others open on Hugging Face.

Downloads over time
Now9.6K→from2K↑373%
1.6K4.5K7.5K10.4K2K on Jul 19.6K on Oct 11JulAugSepOct
Jul 1 → Oct 11 · 54 snapshots · spans 102 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
agpl-3.0
Languages
en th
Tags
transformers safetensors gguf qwen moe fable abliterated sft lora opus reasoning agent

Related

Total size
0 B
Files
3
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-07-09 07:15

Files by quantization

Auxiliary files 3 files 44.3 KB
donate.webp 37.5 KB 9f49908d download
.gitattributes 3.53 KB 3026a75d download
README.md 3.25 KB 4190d0b2 download

README current version from Hugging Face


license: agpl-3.0
language:

  • en
  • th
    tags:
  • qwen
  • moe
  • fable
  • abliterated
  • sft
  • lora
  • opus
  • reasoning
  • gguf
  • agent
  • world-model
  • tool-use
  • conversational
  • trl
  • peft
  • ykai
    base_model:
  • huihui-ai/Huihui-Qwen-AgentWorld-35B-A3B-abliterated
  • huihui-ai/Huihui-Qwen3.6-35B-A3B-abliterated
    datasets:
  • hotdogs/uka-fable-reasoning
  • 11-47/claude_opus_4.8_max_thinking_5k_v2
  • Qwen/AgentWorldBench
    library_name: transformers
    pipeline_tag: text-generation

🚀 Updated July 2026 — Now with AgentWorld + Opus 4.8 Reasoning!

📦 Available Models

Folder Description Size
model_agent/ 🏆 AgentWorld + Fable + Opus 4.8 (recommended) 54 GB
model_chat/ Fable SFT + Opus 4.8 (chat focused) 54 GB

📥 GGUF Downloads

File Format Size Description
GGUF/...AgentWorld-fable-opus4.8.Q4_K_M.gguf Q4_K_M 20 GB Recommended
GGUF/...AgentWorld-fable-opus4.8.f16.gguf F16 65 GB Full precision
GGUF/...Fable-opus4.8-CHAT.Q4_K_M.gguf Q4_K_M 20 GB Chat version
GGUF/...Fable-opus4.8-CHAT.f16.gguf F16 65 GB Chat F16

🧬 LoRA Adapters

Adapter Description Base
adapter_qwen3.6_35b_opus4.8_sft/ 🔥 Opus 4.8 reasoning SFT Huihui 35B
adapter_qwen3.6_35b_fable_sft/ Fable SFT (perfect-v1) Huihui 35B
adapter/ Original Fable SFT Qwen 35B

🚀 Usage

llama.cpp

./llama-cli -m GGUF/Qwen-35B-A3B-abliterated-AgentWorld-fable-opus4.8.Q4_K_M.gguf \
  --temp 0.6 --top-k 25 --top-p 0.9 --min-p 0.1 --repeat-penalty 1.15

Python

from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained(
    "hotdogs/qwen35b-a3b-fable-sft-abliterated", subfolder="model_agent"
)

📊 Training Pipeline

Qwen3.6-35B-A3B-Base (MoE, 256 experts)
  │
  ├── Huihui abliterated (uncensored)
  │
  ├── Fable SFT (perfect-v1, 3,376 rows)
  │
  ├── Opus 4.8 SFT (6,956 rows, Claude reasoning) 🔥
  │
  └── AgentWorld (tool use, environment simulation) 🏆

Built with ❤️ by UKA - 18yo coder & cybersecurity expert


💖 Support / โปรดสนับสนุน

If you find this model useful, please consider supporting my work!
หากคุณคิดว่าโมเดลนี้มีประโยชน์ กรุณาสนับสนุนผลงานของฉันด้วยนะคะ! 🙏

Bitcoin QR — Donate

₿ Bitcoin — BTC:

bc1qf27cyk3vmugcdyv9xdtuv5jwz37863crpj5c9v

Thank you for your support! 🙏✨
ขอบคุณมากๆ สำหรับการสนับสนุนค่า! 💖🤗


Built with ❤️ by UKA — 18-year-old coder & cybersecurity expert

README history 1 version

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

  1. 2026-07-09Super-squash branch 'main' using huggingface_huba44afec3.2 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