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

noneUsername/Phi-3-medium-4k-instruct-abliterated-v3-W8A8-Dynamic-Per-Token

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%2FPhi-3-medium-4k-instruct-abliterated-v3-W8A8-Dynamic-Per-Token"
Response includes
  • classification m1
  • files 16
  • benchmarks 5 entries
  • hub_downloads_all_time 101
  • 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
101
13 last 30d - stable
Likes
0
Model age
23mo ago
created 2024-11-13
Downloads over time
Now108→from6↑1,700%
040791196 on Nov 13, 2024108 on Oct 11Nov '24Feb '25May '25Aug '25Nov '25FebMayAug
Nov 13, 2024 → Oct 11 · 139 snapshots · spans 697 days

Benchmarks

Benchmark Score Source
BBH average 0.5636269295628487 OpenLLM-v2
IFEval instruct 0.6834532374100719 OpenLLM-v2
IFEval-Prompt 0.5804066543438078 OpenLLM-v2
MATH lvl 5 0.14123867069486404 OpenLLM-v2
MMLU-Pro 0.4399933510638298 OpenLLM-v2

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.

failspy/Phi-3-medium-4k-instruct-abliterated-v3↓ 94 this lineage
noneUsername/Phi-3-medium-4k-instruct-abliterated-v3-W8A8-Dynamic-Per-Token↓ 13

Metadata

Tags
safetensors phi3 custom_code base_model:failspy/Phi-3-medium-4k-instruct-abliterated-v3 base_model:finetune:failspy/Phi-3-medium-4k-instruct-abliterated-v3 8-bit region:us

Related

Other by noneUsername
Qwen3-32B-abliterated-awq 3K QwQ-32B-abliterated-AWQ-INT4 26 huihui-ai-phi-4-abliterated-W8A8 20 Mistral-Small-Instruct-2409-abliterated-W8A8-Dynamic-Per-Token 17 Qwen2.5-Coder-14B-Instruct-abliterated-W8A8-Dynamic-Per-Token 17
Same base model
mradermacher/Phi-3-medium-4k-instruct-abliterated-v3-i1-GGUF 736 mradermacher/Phi-3-medium-4k-instruct-abliterated-v3-GGUF 198 tensorblock/Phi-3-medium-4k-instruct-abliterated-v3-GGUF 90 Stark2008/Phi-3-medium-4k-instruct-abliterated-v3-Q3_K_M-GGUF 53 Iambackup/Phi-3-medium-4k-instruct-abliterated-v3-i1-GGUF 12
Total size
13.3 GB
Files
16
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2024-11-13 13:26
Source huggingface.co/noneUsername/Phi-3-medium-4k-instruct-abliterated-v3-W8A8-Dynamic-Per-Token ↗ Base model failspy/Phi-3-medium-4k-instruct-abliterated-v3 →

Files by quantization

Auxiliary files 16 files 13.3 GB
model-00002-of-00003.safetensors 4.62 GB b2bbac25 download
model-00001-of-00003.safetensors 4.49 GB 00671813 download
model-00003-of-00003.safetensors 4.20 GB 2a144e50 download
tokenizer.json 3.45 MB 759de6dd download
tokenizer.model 488 KB 9e556afd download
modeling_phi3.py 72.0 KB bef0f37c download
model.safetensors.index.json 33.7 KB 81b0e688 download
configuration_phi3.py 10.2 KB f4553db2 download
tokenizer_config.json 3.11 KB 2dadb394 download
config.json 1.97 KB 9235456d download
.gitattributes 1.48 KB a6344aac download
README.md 1.12 KB c4c79dd9 download
special_tokens_map.json 569 B 50b4d340 download
recipe.yaml 395 B 2a324207 download
added_tokens.json 293 B c9d3d3a1 download
generation_config.json 172 B 3f586dc4 download

README current version from Hugging Face

base_model:

  • failspy/Phi-3-medium-4k-instruct-abliterated-v3

vllm (pretrained=/root/autodl-tmp/Phi-3-medium-4k-instruct-abliterated-v3,add_bos_token=true,tensor_parallel_size=2,max_model_len=2048,gpu_memory_utilization=0.80,max_num_seqs=2,enforce_eager=True), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: 1

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.804 ± 0.0252
strict-match 5 exact_match ↑ 0.728 ± 0.0282

vllm (pretrained=/root/autodl-tmp/output0.85,add_bos_token=true,tensor_parallel_size=2,max_model_len=2048,gpu_memory_utilization=0.80,max_num_seqs=5), gen_kwargs: (None), limit: 250.0, num_fewshot: 5, batch_size: 5

Tasks Version Filter n-shot Metric Value Stderr
gsm8k 3 flexible-extract 5 exact_match ↑ 0.812 ± 0.0248
strict-match 5 exact_match ↑ 0.764 ± 0.0269

README history 1 version

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

  1. 2024-11-13Create README.md00d2c9c1.1 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