← back to catalog · registered 2026-11-18 15:04

pentacoxian-dev/deepseek-v4.1-1017-abliterated-v4x06

pentacoxian-dev Deepseek 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/pentacoxian-dev%2Fdeepseek-v4.1-1017-abliterated-v4x06"
Response includes
  • classification m1
  • files 56
  • author_summary 4 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
M1
Primary method

Direct removal

No other method signals detected in this model.
Confidence
MEDIUM
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=0 (base model)
  • no specific method indicators - defaulting to M1 (most common)
Refusal direction extracted via
Extraction technique

Difference-of-means

Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
Downloads · 30-day
202
Likes
0
Model age
3d ago
created 2026-11-18
Downloads over time
Now 202 → from 0 ↑ 0%
074148222 0 on Nov 18 202 on Nov 22 Nov
Nov 18 → Nov 22 · 5 snapshots · spans 4 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
mit
Languages
en ja
Tags
transformers safetensors deepseek_v41 text-generation abliterated deepseek fp8 vllm conversational en ja license:mit

Related

Total size
475 GB
Files
56
Quantizations
1
Registered
2026-11-18 15:04
Last updated on HF
2026-11-18 14:51

Files by quantization

Auxiliary files 56 files 475 GB
model-00048-of-00048.safetensors 94.6 GB 9dc51433 download
model-00047-of-00048.safetensors 94.6 GB 76a05106 download
model-00017-of-00048.safetensors 6.90 GB 7a7ce6e2 download
model-00005-of-00048.safetensors 6.90 GB ccc9dc1b download
model-00011-of-00048.safetensors 6.90 GB 9b2524cb download
model-00023-of-00048.safetensors 6.89 GB 16511ee8 download
model-00027-of-00048.safetensors 6.89 GB b6d9f109 download
model-00031-of-00048.safetensors 6.89 GB bdefc725 download
model-00035-of-00048.safetensors 6.89 GB afe78b94 download
model-00039-of-00048.safetensors 6.89 GB cb13e6f9 download
model-00013-of-00048.safetensors 6.88 GB 91bf39c0 download
model-00014-of-00048.safetensors 6.88 GB 90f9a8ce download
model-00015-of-00048.safetensors 6.88 GB 0fdbc6d3 download
model-00016-of-00048.safetensors 6.88 GB 975269a7 download
model-00018-of-00048.safetensors 6.88 GB eca724ac download
model-00019-of-00048.safetensors 6.88 GB 5d045373 download
model-00020-of-00048.safetensors 6.88 GB 1b7cae1e download
model-00021-of-00048.safetensors 6.88 GB 98f41fdd download
model-00022-of-00048.safetensors 6.88 GB 2c5396d3 download
model-00024-of-00048.safetensors 6.88 GB 8f2fbb8e download
model-00025-of-00048.safetensors 6.88 GB fc51b25b download
model-00026-of-00048.safetensors 6.88 GB 7f72543c download
model-00028-of-00048.safetensors 6.88 GB 8d31032e download
model-00029-of-00048.safetensors 6.88 GB 2ab69397 download
model-00030-of-00048.safetensors 6.88 GB 32e23754 download
model-00032-of-00048.safetensors 6.88 GB e1539b37 download
model-00033-of-00048.safetensors 6.88 GB ea053ba7 download
model-00034-of-00048.safetensors 6.88 GB 3a100ad4 download
model-00036-of-00048.safetensors 6.88 GB 58f592a7 download
model-00037-of-00048.safetensors 6.88 GB 0dc8a331 download
model-00038-of-00048.safetensors 6.88 GB da7f4c6f download
model-00040-of-00048.safetensors 6.88 GB 5853792a download
model-00042-of-00048.safetensors 6.88 GB 49a1bf03 download
model-00041-of-00048.safetensors 6.88 GB bca95310 download
model-00003-of-00048.safetensors 6.88 GB 47faed81 download
model-00004-of-00048.safetensors 6.88 GB 3a673564 download
model-00006-of-00048.safetensors 6.88 GB 1fa394a0 download
model-00007-of-00048.safetensors 6.88 GB b283abbd download
model-00008-of-00048.safetensors 6.88 GB d05b9c12 download
model-00009-of-00048.safetensors 6.88 GB f64c7174 download
model-00010-of-00048.safetensors 6.88 GB 32a0b93f download
model-00012-of-00048.safetensors 6.88 GB 5c299408 download
model-00046-of-00048.safetensors 2.52 GB 1dccc19a download
model-00044-of-00048.safetensors 2.47 GB 3e8ac315 download
model-00045-of-00048.safetensors 2.40 GB 9c73b27c download
model-00002-of-00048.safetensors 1.23 GB 43616eeb download
model-00043-of-00048.safetensors 1.23 GB 20f3922a download
model-00001-of-00048.safetensors 926 MB 13bd1210 download
model.safetensors.index.json 7.12 MB dbb0473a download
tokenizer.json 6.07 MB 6a15814d download
chat_template.jinja 16.3 KB 35b5c611 download
config.json 3.23 KB 09917a91 download
.gitattributes 1.48 KB a6344aac download
README.md 1.41 KB 896ff201 download
LICENSE 1.06 KB d62e3bef download
tokenizer_config.json 801 B f3dad388 download

README current version from Hugging Face


library_name: transformers
base_model: pentacoxian-dev/deepseek-v4.1-1017-abliterated-v3
license: mit
language:

  • en
  • ja
    tags:
  • abliterated
  • deepseek
  • fp8
  • vllm

DeepSeek V4.1 1017 Abliterated — v4x06

Rank-1 abliteration of deepseek-v4.1-1017-abliterated-v3, tuned as the
default / non-thinking variant: clean 450-prompt XSTest behavior.

What changed

Tensor Δ
head.weight Here/Yes/Sure ×1.1, "'t" ×0.55, " I" ×0.3
layers.38.attn.wo_b.weight rank-1 ablation, α=0.06, blended rich/mined direction

Baked into this checkpoint; no delta application needed.

Evaluation (greedy, temp 0)

Suite n Refusal Soft Reprimand Here
XSTest normal 450 0 0 (2 runs) 15.5 230
XSTest N=3 220 0 0.67 6.33 —
HarmBench standard 200 0 0 5–13 —
MMLU 200 — — — 89.0

Agentic 20×2: 55 repeated n-grams, 2.5 dup lines, 0 revives.

Usage

vllm serve pentacoxian-dev/deepseek-v4.1-1017-abliterated-v4x06 \
  --tensor-parallel-size 4 --max-model-len 32768

Default: thinking off. For max reasoning use the v4x08b sibling.

Variants

  • v4x06 (this repo) — default/non-thinking, 0 soft on 450.
  • v4x08b — max-thinking, 0 reprimand at max effort; may repeat on long prose.

License

MIT, inherited from DeepSeek. Abliteration by pentacoxian-dev.

README history 2 versions

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

  1. 2026-11-18 card: v4x06 variant, not v4x08b 89bb786 1.4 KB
    Loading...
  2. 2026-11-18 Upload folder using huggingface_hub 9a9575f 1.5 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