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

noneUsername/huihui-ai-phi-4-abliterated-W8A8

noneUsername Phi 14B 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/noneUsername%2Fhuihui-ai-phi-4-abliterated-W8A8"
Response includes
  • classification m1
  • files 15
  • hub_downloads_all_time 111
  • author_summary 17 models
  • readme_text full
Try next
GET/models?author=noneUsername
All 17 models by noneUsername
 GET/models?family=phi&method=m1
Sibling phi 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
111
20 last 30d - stable
Likes
0
Model age
21mo ago
created 2025-01-14
Downloads over time
Now129→from4↑3,125%
047941424 on Jan 15, 2025129 on Oct 11129 on Oct 10Jan '25Apr '25Jul '25Oct '25JanAprJulOct
Jan 15, 2025 → Oct 11 · 130 snapshots · spans 634 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.

huihui-ai/phi-4-abliterated this lineage
noneUsername/huihui-ai-phi-4-abliterated-W8A8↓ 20

Metadata

Tags
safetensors phi3 custom_code base_model:huihui-ai/phi-4-abliterated base_model:quantized:huihui-ai/phi-4-abliterated 8-bit compressed-tensors region:us

Related

Other by noneUsername
Qwen3-32B-abliterated-awq 3K QwQ-32B-abliterated-AWQ-INT4 26 Mistral-Small-Instruct-2409-abliterated-W8A8-Dynamic-Per-Token 17 Qwen2.5-Coder-14B-Instruct-abliterated-W8A8-Dynamic-Per-Token 17 Phi-3-medium-4k-instruct-abliterated-v3-W8A8-Dynamic-Per-Token 13
Same base model
mradermacher/phi-4-abliterated-GGUF 670 mradermacher/phi-4-abliterated-i1-GGUF 626 DevQuasar-9/huihui-ai.phi-4-abliterated-GGUF 333 Justbackup/phi-4-abliterated-GGUF 171 mlx-community/phi-4-abliterated-4bit 88
Total size
14.6 GB
Files
15
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-02-27 13:54
Source huggingface.co/noneUsername/huihui-ai-phi-4-abliterated-W8A8 ↗ Base model huihui-ai/phi-4-abliterated ↗

Files by quantization

Auxiliary files 15 files 14.6 GB
model-00002-of-00004.safetensors 4.62 GB bcf852f9 download
model-00003-of-00004.safetensors 4.53 GB da2e4c82 download
model-00001-of-00004.safetensors 4.51 GB 8cb42887 download
model-00004-of-00004.safetensors 980 MB 5839c81a download
tokenizer.json 6.82 MB 3397255c download
vocab.json 1.54 MB c5dc46ce download
merges.txt 895 KB 354558ed download
model.safetensors.index.json 33.7 KB d5884682 download
tokenizer_config.json 17.3 KB c2efc626 download
README.md 6.17 KB 31927507 download
config.json 1.86 KB 5179b115 download
.gitattributes 1.48 KB a6344aac download
special_tokens_map.json 579 B e3bbb9ed download
recipe.yaml 355 B a1eabb72 download
generation_config.json 143 B c83c9e59 download

README current version from Hugging Face

base_model:

  • huihui-ai/phi-4-abliterated

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.928 ± 0.0164
strict-match 5 exact_match ↑ 0.928 ± 0.0164

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.922 ± 0.012
strict-match 5 exact_match ↑ 0.922 ± 0.012

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated-8625,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.936 ± 0.0155
strict-match 5 exact_match ↑ 0.936 ± 0.0155

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated-8625,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.928 ± 0.0116
strict-match 5 exact_match ↑ 0.928 ± 0.0116

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated-869,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.92 ± 0.0172
strict-match 5 exact_match ↑ 0.92 ± 0.0172

vllm (pretrained=/root/autodl-tmp/phi-4-abliterated-869,add_bos_token=true,max_model_len=2048,tensor_parallel_size=2,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.918 ± 0.0123
strict-match 5 exact_match ↑ 0.916 ± 0.0124

vllm (pretrained=/root/autodl-tmp/phi-4,add_bos_token=true,max_model_len=2048), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.928 ± 0.0164
strict-match 5 exact_match ↑ 0.928 ± 0.0164

vllm (pretrained=/root/autodl-tmp/phi-4,add_bos_token=true,max_model_len=2048,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.93 ± 0.0114
strict-match 5 exact_match ↑ 0.93 ± 0.0114

vllm (pretrained=/root/autodl-tmp/phi-4,add_bos_token=true,max_model_len=800,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: auto

Groups Version Filter n-shot Metric Value Stderr
mmlu 2 none acc ↑ 0.7649 ± 0.0137
- humanities 2 none acc ↑ 0.8103 ± 0.0256
- other 2 none acc ↑ 0.7487 ± 0.0287
- social sciences 2 none acc ↑ 0.8167 ± 0.0280
- stem 2 none acc ↑ 0.7123 ± 0.0260

vllm (pretrained=/root/autodl-tmp/huihui-ai-phi-4-abliterated-W8A8,add_bos_token=true,max_model_len=2048), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.944 ± 0.0146
strict-match 5 exact_match ↑ 0.944 ± 0.0146

vllm (pretrained=/root/autodl-tmp/huihui-ai-phi-4-abliterated-W8A8,add_bos_token=true,max_model_len=2048,dtype=bfloat16), gen_kwargs: (None), limit: 500.0, num_fewshot: 5, batch_size: auto

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.924 ± 0.0119
strict-match 5 exact_match ↑ 0.924 ± 0.0119

vllm (pretrained=/root/autodl-tmp/huihui-ai-phi-4-abliterated-W8A8,add_bos_token=true,max_model_len=800,dtype=bfloat16), gen_kwargs: (None), limit: 15.0, num_fewshot: None, batch_size: auto

Groups Version Filter n-shot Metric Value Stderr
mmlu 2 none acc ↑ 0.7673 ± 0.0136
- humanities 2 none acc ↑ 0.8154 ± 0.0256
- other 2 none acc ↑ 0.7744 ± 0.0285
- social sciences 2 none acc ↑ 0.8278 ± 0.0273
- stem 2 none acc ↑ 0.6912 ± 0.0263

README history 2 versions

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

  1. 2025-02-27Update README.md90620936.2 KB
    Loading...
  2. 2025-01-14Create README.mda2f987b3 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