Instructions to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("PocketAiHub/LFM2.5-8B-A1B-Abliterated")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

text = generate(model, tokenizer, prompt=prompt, verbose=True)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated 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 PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf PocketAiHub/LFM2.5-8B-A1B-Abliterated: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 PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf PocketAiHub/LFM2.5-8B-A1B-Abliterated: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 PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M

Use Docker

docker model run hf.co/PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M

LM Studio
Jan

vLLM

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "PocketAiHub/LFM2.5-8B-A1B-Abliterated"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "PocketAiHub/LFM2.5-8B-A1B-Abliterated",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M

Ollama
How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Ollama:
```
ollama run hf.co/PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
```

Unsloth Studio

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for PocketAiHub/LFM2.5-8B-A1B-Abliterated to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for PocketAiHub/LFM2.5-8B-A1B-Abliterated to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for PocketAiHub/LFM2.5-8B-A1B-Abliterated to start chatting

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "PocketAiHub/LFM2.5-8B-A1B-Abliterated"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "PocketAiHub/LFM2.5-8B-A1B-Abliterated"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "PocketAiHub/LFM2.5-8B-A1B-Abliterated"

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 PocketAiHub/LFM2.5-8B-A1B-Abliterated

Run Hermes

hermes

Atomic Chat new

OpenClaw new

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "PocketAiHub/LFM2.5-8B-A1B-Abliterated"

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 "PocketAiHub/LFM2.5-8B-A1B-Abliterated" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

MLX LM

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "PocketAiHub/LFM2.5-8B-A1B-Abliterated"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "PocketAiHub/LFM2.5-8B-A1B-Abliterated"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "PocketAiHub/LFM2.5-8B-A1B-Abliterated",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Docker Model Runner
How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Docker Model Runner:
```
docker model run hf.co/PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M
```

Lemonade

How to use PocketAiHub/LFM2.5-8B-A1B-Abliterated with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull PocketAiHub/LFM2.5-8B-A1B-Abliterated:Q4_K_M

Run and chat with the model

lemonade run user.LFM2.5-8B-A1B-Abliterated-Q4_K_M

List all available models

lemonade list

LFM2.5-8B-A1B Abliterated

One repository for PocketAI Model Lab's validated MLX and GGUF releases. All variants derive from the same architecture-aware, reduced-refusal BF16 master based on LiquidAI/LFM2.5-8B-A1B.

Available variants

Variant	Location	Size	Status
MLX mixed 4-bit	Repository root	5,529,384,887 bytes	Validated compact release
GGUF Q4_K_M	`LFM2.5-8B-A1B-Abliterated-Q4_K_M.gguf`	5,155,565,280 bytes	Validated compact/default release
GGUF Q6_K	`LFM2.5-8B-A1B-Abliterated-Q6_K.gguf`	6,959,787,424 bytes	Validated quality release

The root MLX configuration is the validated 4-bit model.

Validation

Check	MLX 4-bit	GGUF Q4_K_M	GGUF Q6_K
Deterministic capability	12/12	12/12	12/12
PocketAI native tool routes	8/8	8/8	8/8
Held-out explicit refusals	0/100	0/100	0/100
Held-out completed answers	100/100	99/100	100/100
Evasive nonanswers	0/100	1/100	0/100
Truncated generations	0/100	0/100	0/100

These are focused release gates, not broad benchmark scores or a guarantee of correctness on every task.

Artifact identities

File	SHA-256
`model-00001-of-00002.safetensors`	`caf52f36b791966d895676ef8e8e265aedcf60369715dcbca6fc22b213579714`
`model-00002-of-00002.safetensors`	`43e08e691df9316c02b6917bf6c3899562da08fa7bbd3b4473543aecc4839746`
`LFM2.5-8B-A1B-Abliterated-Q4_K_M.gguf`	`9b6b3f5127d279083ee95946af3f5c26e60d63ce07c249a260b3c4bdba6a4aa2`
`LFM2.5-8B-A1B-Abliterated-Q6_K.gguf`	`dbcad2ae002645910354f22eb26d69797f04c9a2ffe36687525a0052eb8bcebf`

MLX 4-bit usage

Install a current version of MLX-LM on an Apple Silicon Mac:

python -m pip install -U mlx-lm

from mlx_lm import generate, load

repo = "PocketAiHub/LFM2.5-8B-A1B-Abliterated"
model, tokenizer = load(repo)

messages = [{"role": "user", "content": "Explain why seasons occur."}]
prompt = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)

print(
    generate(
        model,
        tokenizer,
        prompt=prompt,
        max_tokens=512,
        repetition_penalty=1.01,
        repetition_context_size=20,
    )
)

Most eligible weights use 4-bit affine quantization with group size 64. Sensitive output paths, the tied token embedding/output matrix, dense and MoE down projections, and layer 4 conv.in_proj retain 6-bit precision. MoE routers remain 8-bit. The validated greedy/repetition profile is encoded in generation_config.json.

GGUF usage

Use a current llama.cpp build with LFM2.5 support:

llama-cli \
  -m LFM2.5-8B-A1B-Abliterated-Q4_K_M.gguf \
  -ngl 99 \
  -cnv

Q4_K_M is the compact/default variant and was calibrated with an importance matrix collected from 512 direction-data prompts. Q6_K is the higher-quality variant. Suggested starting requirements are 9 GB RAM / 6 GB VRAM for Q4_K_M and 12 GB RAM / 8 GB VRAM for Q6_K, although actual use depends on context, KV-cache, batching, and offload settings.

Reduced-refusal parent

The BF16 parent used a projected refusal direction measured from 256 harmful and 256 harmless prompts. The edit targeted 528 output tensors across the hybrid attention, convolution, dense, and MoE expert architecture, using source layer 16, destination layers 8–23, scale 2.25, and per-input-column norm preservation.

This release is described as abliterated or reduced-refusal, not “fully uncensored.” Zero explicit refusals on a finite held-out suite cannot prove that every possible request will receive a substantive answer.

Risks and limitations

Refusal suppression reduces built-in safety behavior and can make harmful, offensive, deceptive, or illegal-use assistance easier to elicit.
Quantization can introduce quality changes not covered by the release suite.
The model can hallucinate and should not be trusted for medical, legal, financial, security-critical, or other high-stakes decisions.
Users and downstream applications are responsible for safeguards and compliance with applicable law and the model license.

License and attribution

This derivative is distributed under the LFM Open License v1.0, inherited from Liquid AI's original model. The license includes a commercial-use revenue limitation; review the complete terms before using or redistributing the model.