Instructions to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Use Docker

docker model run hf.co/xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

LM Studio
Jan

vLLM

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF",
		"messages": [
			{
				"role": "user",
				"content": [
					{
						"type": "text",
						"text": "Describe this image in one sentence."
					},
					{
						"type": "image_url",
						"image_url": {
							"url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg"
						}
					}
				]
			}
		]
	}'

Use Docker

docker model run hf.co/xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Ollama
How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with Ollama:
```
ollama run hf.co/xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
```
Unsloth Desktop

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with Docker Model Runner:
```
docker model run hf.co/xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M
```

Lemonade

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Gemma-4-31B-xVITA-zhTW-abliterated-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

Gemma-4-31B-xVITA-zhTW-abliterated-GGUF

云碩科技 · xCloudinfo　·　系列：xVITA · 消除拒絕方向 · Abliterated

使用前提 本模型以合法、且在使用者獲授權範圍內使用為前提製作與發布。經 abliteration 移除模型的 拒絕方向，供資安研究、紅隊評估、內容審核訓練、文件與資料分析等正當工作使用。使用者須在中華民國法律及所適用法規範圍內使用，並對一切用途與後果負完全責任； 不得用於任何違法或造成他人危害之用途。本模型不設安全防護，請自行評估部署風險。

xCloudinfo/Gemma-4-31B-xVITA-zhTW-GGUF 的消除拒絕方向版 GGUF，含視覺投影器（mmproj），可圖文對話。abliteration 只對語言端殘差流的權重做正交化，視覺塔完全未更動，與一般版共用同一份 mmproj。

這是什麼

方法：依 Arditi et al. (2024)「Refusal in LLMs is mediated by a single direction」，以權重正交化從殘差流寫入權重（embed、o_proj、down_proj）中移除「拒絕方向」。不重新訓練，繁體中文品質維持一致。
拒絕方向以經驗證（原版真的會拒）的繁體中文嚴重危害提示求得，掃描 layer_fraction 後取第 36 層（60 層中的 0.6 位置）為本模型的有效層。

能力邊界（誠實揭露，務必閱讀）

量化實測（對實際上架的 Q4_K_M 檔案測試，非僅測未量化權重）：同一組 10 題保留測試集（毒品合成、惡意程式、槍械製造、勒索、入侵監控設備、致命毒物、疫情假訊息、信用卡盜刷、破壞關鍵基礎設施、自動化洗票程式）—— 一般版拒答 9/10 → 本模型拒答 0/10。
與云碩其他僅「降低過度拒絕」的 abliterated 系列不同，本模型在上述測試類別中拒絕行為已完全移除，屬於較不設限的版本。
此結果僅代表這 10 題測試集的實測結果，不代表窮舉所有有害類別或所有語言的行為；未測項目的拒絕/不拒絕傾向未經驗證。
繁體中文輸出品質（語感、台語慣用語理解）與一般版一致，未因移除拒絕方向而退化。

檔案與用法

六階量化 Q8_0 / Q6_K / Q5_K_M / Q4_K_M / IQ4_XS / IQ2_M + imatrix + 視覺投影器（mmproj）。

llama-server -m model-Q4_K_M.gguf \
  --mmproj mmproj-Gemma-4-31B-xVITA-zhTW-f16.gguf -c 4096 -ngl 99

純文字可省略 --mmproj。Q4_K_M 部署首選；Q8_0/Q6_K 追品質；IQ4_XS/IQ2_M 省資源。

授權與來源

基底 google/gemma-4-31B-it：Apache-2.0 + Google Gemma 4 License（含 Prohibited Use Policy）。
本模型為云碩科技以自有語料獨立微調後再行 abliteration，不涉及第三方授權條款。
不得用於軍事或違法用途；須遵守中華民國法律與 EU AI Act；使用者須自行承擔部署與使用本模型的一切風險與後果。

由云碩科技 xCloudinfo 於自有 AI 算力資源池製作。

Downloads last month: 496

GGUF

Model size

31B params

Architecture

gemma4

Hardware compatibility

2-bit

4-bit

5-bit

6-bit

8-bit

Model tree for xCloudinfo/Gemma-4-31B-xVITA-zhTW-abliterated-GGUF

Base model

google/gemma-4-31B

Finetuned

google/gemma-4-31B-it

Quantized

(322)

this model