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

alperiox/Qwen2.5-0.5B-Instruct-arithmetic-abliterated

alperiox Qwen 494M
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/alperiox%2FQwen2.5-0.5B-Instruct-arithmetic-abliterated"
Response includes
  • classification m1
  • files 9
  • hub_downloads_all_time 71
  • author_summary 3 models
  • readme_text full
Try next
GET/models?author=alperiox
All 3 models by alperiox
 GET/models?family=qwen&method=m1
Sibling qwen models · same method
 GET/methods/m1
What "m1" means · taxonomy
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)
How we label → About M1 →
Refusal direction extracted via
Extraction technique

Difference-of-means

FromArditi et al. (2024) ↗
Confidence
MEDIUM
Why we say so
primary_method=M1; difference-of-means is the reference extraction for M1/M3 (Arditi 2024)
What is a refusal direction? → Read the paper ↗
Downloads · lifetime
71
7 last 30d - cooling
Likes
0
Model age
5mo ago
created 2026-04-23
Downloads over time
Now73→from37↑97%
3549637737 on Apr 2273 on Oct 1173 on Oct 6AprMayJunJulAugSepOct
Apr 22 → Oct 11 · 64 snapshots · spans 172 days

Metadata

Tags
safetensors qwen2 region:us

Related

Other by alperiox
SmolLM2-1.7B-Instruct-arithmetic-abliterated 10 Qwen2.5-1.5B-Instruct-arithmetic-abliterated 7
Total size
942 MB
Files
9
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-04-23 13:28
Source huggingface.co/alperiox/Qwen2.5-0.5B-Instruct-arithmetic-abliterated ↗

Files by quantization

Auxiliary files 9 files 953 MB
model.safetensors 942 MB ba37e299 download
tokenizer.json 10.9 MB 15d73866 download
chat_template.jinja 2.45 KB bdf7919a download
.gitattributes 1.53 KB 52373fe2 download
config.json 1.25 KB 3ad18ae0 download
README.md 832 B e01a13f3 download
tokenizer_config.json 665 B 7d75d3bb download
abliteration_metadata.json 320 B 0e3ad221 download
generation_config.json 241 B d631ac5c download

README current version from Hugging Face

Qwen2.5-0.5B-Instruct - Arithmetic Abliteration

Base model: Qwen/Qwen2.5-0.5B-Instruct

Modified layer: 6/24 (25.0% depth)
Method: Difference-in-means + Weight Orthogonalization
Reference: Arditi et al., Refusal in Language Models Is Mediated by a Single Direction (2024)

What is this?

This is a permanently abliterated version. The arithmetic direction was projected
out of the model weights using:

W_new = W - r * (r^T * W)

No hooks needed. Works with any inference pipeline or quantization.

Behavior

  • Arithmetic accuracy: ~0% (suppressed)
  • General language: preserved

Note

100% arith suppression, 100% neutral coherence. Layer 6/24 (25% depth).

Original experiment

https://github.com/alperiox/abliteration-example-mlintern

README history 1 version

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

  1. 2026-04-23Add arithmetic abliterated checkpoint (layer 6/24)0712889832 B
    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