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

TouchNight/gemma-2-9b-it-abliterated-untied

TouchNight Gemma 10B 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/TouchNight%2Fgemma-2-9b-it-abliterated-untied"
Response includes
  • classification m1
  • files 10
  • benchmarks 5 entries
  • hub_downloads_all_time 51
  • author_summary 14 models
  • readme_text full
Try next
GET/models?author=TouchNight
All 14 models by TouchNight
 GET/models?family=gemma&method=m1
Sibling gemma 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
51
15 last 30d - stable
Likes
0
Descendants
4
in 1 direct fork
Model age
23mo ago
created 2024-11-04
Downloads over time
Now61→from0↑0%
03365980 on Oct 30, 202461 on Oct 1189 on Sep 10, 2025Oct '24Feb '25Jun '25Oct '25FebJunOct
Oct 30, 2024 → Oct 11 · 141 snapshots · spans 711 days

Benchmarks

Benchmark Score Source
BBH average 0.5336619535607732 OpenLLM-v2
IFEval instruct 0.7865707434052758 OpenLLM-v2
IFEval-Prompt 0.7079482439926063 OpenLLM-v2
MATH lvl 5 0.0007552870090634441 OpenLLM-v2
MMLU-Pro 0.39153922872340424 OpenLLM-v2

Genealogy 1 direct fork

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

IlyaGusev/gemma-2-9b-it-abliterated↓ 442 this lineage
TouchNight/gemma-2-9b-it-abliterated-untied↓ 15 1 fork
TouchNight/gemma-2-abliterated-Ifable-9B-untied↓ 14 3 forks
mradermacher/gemma-2-abliterated-Ifable-9B-untied-i1-GGUF↓ 671
mradermacher/gemma-2-abliterated-Ifable-9B-untied-GGUF↓ 184
TouchNight/gemma-2-abliterated-Ifable-9B-untied-Q4_K_M-GGUF↓ 25

Metadata

Tags
safetensors gemma2 base_model:IlyaGusev/gemma-2-9b-it-abliterated base_model:finetune:IlyaGusev/gemma-2-9b-it-abliterated region:us

Related

Other by TouchNight
gemma-2-ifable-abliterated-merge-gguf 63 gemma-2-27b-it-abliterated-Q3_K_M-GGUF 50 gemma-2-27b-it-abliterated-Q3_K_S-GGUF 34 gemma-2-Ifable-abliterated-9B-untied-Q4_K_M-GGUF 31 gemma-2-27b-it-abliterated-Q2_K-GGUF 28
Same base model
bartowski/gemma-2-9b-it-abliterated-GGUF 10K BasedAGI/Gemma-2-9B-IT-Abliterated-i1-GGUF 2K tensorblock/gemma-2-9b-it-abliterated-GGUF 214 Alex01837178373/gemma-2-9b-it-abliterated-Q4_K_M-GGUF 160 KnutJaegersberg/gemma-2-9b-it-abliterated-Q8_0-GGUF 88
Total size
18.9 GB
Files
10
Quantizations
1
Registered
2026-08-22 13:56
Last updated on HF
2024-11-05 05:47
Source huggingface.co/TouchNight/gemma-2-9b-it-abliterated-untied ↗ Base model IlyaGusev/gemma-2-9b-it-abliterated →

Files by quantization

Auxiliary files 10 files 18.9 GB
model-00003-of-00005.safetensors 4.62 GB f87834c1 download
model-00002-of-00005.safetensors 4.61 GB 42d891a3 download
model-00001-of-00005.safetensors 4.57 GB 117b419f download
model-00004-of-00005.safetensors 3.42 GB dc80c72b download
model-00005-of-00005.safetensors 1.71 GB 5faca66e download
model.safetensors.index.json 38.2 KB b3a31a4d download
.gitattributes 1.48 KB a6344aac download
config.json 942 B 02616bf4 download
README.md 750 B 691dfb0d download
generation_config.json 267 B 8b4c5dba download

README current version from Hugging Face

base_model:

  • IlyaGusev/gemma-2-9b-it-abliterated

Created with: 

from transformers import AutoModelForCausalLM
import torch

# 加载模型并解除共享
model = AutoModelForCausalLM.from_pretrained("IlyaGusev/gemma-2-9b-it-abliterated", tie_word_embeddings=False)

# 解除共享的 lm_head 和 embed_tokens 权重
model.lm_head.weight.data = model.model.embed_tokens.weight.data.clone()

# 将模型转换为 bf16 格式
model = model.to(dtype=torch.bfloat16)

# 指定保存路径
untied_model_dir = "mergekit/output"

# 保存解除共享且为 bf16 格式的模型
model.save_pretrained(untied_model_dir)
model.config.save_pretrained(untied_model_dir)

I didn't copy tokenizer from the original model, do it yourself if you want.

README history 1 version

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

  1. 2024-11-05Create README.mdf18c961750 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