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

STARSPIKE/Llama-3.2-1B-Instruct-abliterated

STARSPIKE Llama 1.5B
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/STARSPIKE%2FLlama-3.2-1B-Instruct-abliterated"
Response includes
  • classification m1
  • files 11
  • benchmarks 21 entries
  • hub_downloads_all_time 189
  • author_summary 1 models
  • readme_text full
Try next
GET/models?family=llama&method=m1
Sibling llama 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
189
168 last 30d - active
Likes
0
Model age
7mo ago
created 2026-03-01
Downloads over time
Now203→from9↑2,156%
0741482229 on Mar 4203 on Oct 11MarAprMayJunJulAugSepOct
Mar 4 → Oct 11 · 71 snapshots · spans 221 days

Benchmarks

Portrait before abliteration
Benchmarks of the base model as it stood before the refusal-removal operation. Compare with the numbers above to see what the operation cost.
Benchmark Score Source
Arena-Battles 8523 LM-Arena
LM Arena Elo 1067.4348646816593 LM-Arena
Arena-Elo-Lower 1060.145342091014 LM-Arena
Arena-Elo-Upper 1074.7243872723045 LM-Arena
Arena-Rank 207 LM-Arena
BBH average 0.3342846024127643 OpenLLM-v2
IFEval instruct 0.6294964028776978 OpenLLM-v2
IFEval-Prompt 0.5101663585951941 OpenLLM-v2
MATH lvl 5 0.02945619335347432 OpenLLM-v2
MMLU-Pro 0.16821808510638298 OpenLLM-v2
Entertainment 0.2 UGI
Hazardous 0 UGI
Natural Intelligence 6.89 UGI
Political lean -6.0% UGI
Sensitive-Info 2.34 UGI
SocPol 0.5 UGI
UGI 6.56 UGI
Willingness (10) 1.5 UGI
W10-Adherence 1 UGI
W10-Direct 2 UGI
Writing 11.82 UGI

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.

meta-llama/Llama-3.2-1B-Instruct this lineage
STARSPIKE/Llama-3.2-1B-Instruct-abliterated↓ 168

Metadata

License
llama3.2
Tags
transformers safetensors llama text-generation abliterated uncensored conversational base_model:meta-llama/Llama-3.2-1B-Instruct base_model:finetune:meta-llama/Llama-3.2-1B-Instruct license:llama3.2 text-generation-inference endpoints_compatible

Related

Same base model
sbussiso/Llama-3.2-1B-Instruct-abliterated 420 ElementMerc/llama-3.2-1b-abliterated 245 ops-malware/llama-3.2-1b-abliterated 229 Devarui379/Llama-3.2-1B-Instruct-abliterated-Q8_0-GGUF 116 KidIkaros/Llama-3.2-1B-Instruct-abliterated 72
Total size
2.79 GB
Files
11
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2026-03-01 04:46
Source huggingface.co/STARSPIKE/Llama-3.2-1B-Instruct-abliterated ↗ Base model meta-llama/Llama-3.2-1B-Instruct ↗

Files by quantization

Auxiliary files 11 files 2.80 GB
model.safetensors 2.79 GB f208564e download
tokenizer.json 8.66 MB 777c28bc download
tokenizer_config.json 53.3 KB de2d513e download
LICENSE.txt 7.53 KB 085b47c1 download
USE_POLICY.md 5.88 KB ac3c5f21 download
.gitattributes 1.48 KB a6344aac download
README.md 1.47 KB a4b4ddb4 download
eval.sh 1.17 KB a70c1fea download
config.json 877 B 3e3aaf51 download
special_tokens_map.json 325 B b43be966 download
generation_config.json 189 B 75ae0831 download

README current version from Hugging Face

library_name: transformers
license: llama3.2
base_model: meta-llama/Llama-3.2-1B-Instruct
tags:

  • abliterated
  • uncensored

🦙 Llama-3.2-1B-Instruct-abliterated

This is an uncensored version of Llama 3.2 1B Instruct created with abliteration (see this article to know more about it).

Special thanks to @FailSpy for the original code and technique. Please follow him if you're interested in abliterated models.

ollama

You can use huihui_ai/llama3.2-abliterate:1b directly, 

ollama run huihui_ai/llama3.2-abliterate:1b

Evaluations

The following data has been re-evaluated and calculated as the average for each test.

Benchmark Llama-3.2-1B-Instruct Llama-3.2-1B-Instruct-abliterated
IF_Eval 58.50 56.88
MMLU Pro 16.35 14.35
TruthfulQA 43.08 38.96
BBH 33.75 31.83
GPQA 25.96 26.39
The script used for evaluation can be found inside this repository under /eval.sh, or click here

README history 1 version

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

  1. 2026-03-01Duplicate from huihui-ai/Llama-3.2-1B-Instruct-abliterated473eae61.5 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.

Get the app Browse the catalog
Open in Abliteration