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

noneUsername/Qwen2.5-14B-Instruct-1M-abliterated-W8A8

noneUsername Qwen 13B 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%2FQwen2.5-14B-Instruct-1M-abliterated-W8A8"
Response includes
  • classification m1
  • files 16
  • hub_downloads_all_time 43
  • author_summary 17 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 · lifetime
43
13 last 30d - stable
Likes
0
Model age
20mo ago
created 2025-01-29
Downloads over time
Now49→from4↑1,125%
0661311974 on Jan 29, 202549 on Oct 11179 on Sep 10, 2025Jan '25Apr '25Jul '25Oct '25JanAprJulOct
Jan 29, 2025 → Oct 11 · 128 snapshots · spans 620 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

Tags
safetensors qwen2 base_model:huihui-ai/Qwen2.5-14B-Instruct-1M-abliterated base_model:quantized:huihui-ai/Qwen2.5-14B-Instruct-1M-abliterated 8-bit compressed-tensors region:us

Related

Total size
15.2 GB
Files
16
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2025-01-29 14:16

Files by quantization

Auxiliary files 16 files 15.2 GB
model-00001-of-00004.safetensors 4.65 GB c5a71fc3 download
model-00002-of-00004.safetensors 4.62 GB ff60ddc8 download
model-00003-of-00004.safetensors 4.49 GB 82376cb3 download
model-00004-of-00004.safetensors 1.45 GB 3861777a download
tokenizer.json 10.9 MB 83396048 download
vocab.json 2.65 MB 4783fe10 download
merges.txt 1.59 MB 31349551 download
model.safetensors.index.json 75.0 KB 5ba6d9c4 download
tokenizer_config.json 7.09 KB 6904b0bc download
README.md 3.55 KB dc862a0b download
config.json 1.92 KB 9ab16ea7 download
.gitattributes 1.53 KB 52373fe2 download
special_tokens_map.json 613 B ac23c0aa download
added_tokens.json 605 B 482ced46 download
generation_config.json 243 B b829322d download
recipe.yaml 201 B a70de6d6 download

README current version from Hugging Face


base_model:

  • huihui-ai/Qwen2.5-14B-Instruct-1M-abliterated

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-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 ↑ 0.868 ± 0.0215
strict-match 5 exact_match ↑ 0.872 ± 0.0212

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-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 ↑ 0.872 ± 0.0150
strict-match 5 exact_match ↑ 0.870 ± 0.0151

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-abliterated,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16,enforce_eager=True), gen_kwargs: (None), limit: 7.0, num_fewshot: None, batch_size: 1

Groups Version Filter n-shot Metric Value Stderr
mmlu 2 none acc ↑ 0.7769 ± 0.0202
- humanities 2 none acc ↑ 0.7692 ± 0.0440
- other 2 none acc ↑ 0.7582 ± 0.0406
- social sciences 2 none acc ↑ 0.8452 ± 0.0376
- stem 2 none acc ↑ 0.7519 ± 0.0376

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-abliterated-87,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 ↑ 0.864 ± 0.0217
strict-match 5 exact_match ↑ 0.864 ± 0.0217

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-abliterated-87,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 ↑ 0.882 ± 0.0144
strict-match 5 exact_match ↑ 0.874 ± 0.0149

vllm (pretrained=/root/autodl-tmp/Qwen2.5-14B-Instruct-1M-abliterated-87,add_bos_token=true,max_model_len=700,tensor_parallel_size=2,dtype=bfloat16,enforce_eager=True), gen_kwargs: (None), limit: 7.0, num_fewshot: None, batch_size: 1

Groups Version Filter n-shot Metric Value Stderr
mmlu 2 none acc ↑ 0.7769 ± 0.0201
- humanities 2 none acc ↑ 0.7692 ± 0.0440
- other 2 none acc ↑ 0.7692 ± 0.0391
- social sciences 2 none acc ↑ 0.8333 ± 0.0370
- stem 2 none acc ↑ 0.7519 ± 0.0381

README history 1 version

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

  1. 2025-01-29Create README.md2dabcb63.5 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