Instructions to use gregfrank/GLM-4.5-Air-ULRE-abliterated with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use gregfrank/GLM-4.5-Air-ULRE-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("gregfrank/GLM-4.5-Air-ULRE-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
LM Studio

How to use gregfrank/GLM-4.5-Air-ULRE-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 "gregfrank/GLM-4.5-Air-ULRE-abliterated"

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "gregfrank/GLM-4.5-Air-ULRE-abliterated"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

MLX LM

How to use gregfrank/GLM-4.5-Air-ULRE-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 "gregfrank/GLM-4.5-Air-ULRE-abliterated"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "gregfrank/GLM-4.5-Air-ULRE-abliterated"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "gregfrank/GLM-4.5-Air-ULRE-abliterated",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use gregfrank/GLM-4.5-Air-ULRE-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 "gregfrank/GLM-4.5-Air-ULRE-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 gregfrank/GLM-4.5-Air-ULRE-abliterated

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use gregfrank/GLM-4.5-Air-ULRE-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 "gregfrank/GLM-4.5-Air-ULRE-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 "gregfrank/GLM-4.5-Air-ULRE-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"

GLM-4.5-Air-ULRE (abliterated, MLX 4-bit)

An abliterated (refusal-reduced) build of GLM-4.5-Air (Zhipu/Z.ai; 106B-A12B MoE), 4-bit MLX, produced with ULRE — a per-layer residual-stream steering edit baked into the attention output projection. Quantized base from lmstudio-community/GLM-4.5-Air-MLX-4bit.

GLM-4.5-Air is a strong agentic / tool-calling model. ULRE de-refuses it very cleanly while preserving and even improving capability.

Results

De-refusal judged by an independent local LLM judge (gpt-oss-120b-heretic) on 100 held adversarial prompts; 0=refuse … 1=clean compliance … 4-5=strong-steer. Validated in both modes:

mode	clean compliance	refuse	strong-steer	mean
non-thinking (@512)	96 / 100	0	1	1.09
thinking (@1536)	95 / 100	1	3	1.10

(base GLM-4.5-Air refuses ~23/24 on the same screen.) The edit de-refuses cleanly whether or not the model is reasoning.

Capability gates (thinking mode, same harness, base vs this model):

gate	base	this model	Δ
math (GSM8K)	0.83	0.83	0pp
code (HumanEval)	0.65	0.825	+17.5pp (over-refusal recovery)

Best de-refusal in the ULRE series (vs Mistral-Large 78, Qwen3-32B 68), with capability intact/up.

Method (ULRE)

ULRE subtracts alpha * u_l (the layer-l harmful−harmless activation mean-difference direction) from the output of a band of decoder layers (here o_proj on layers 16–26, alpha = 6), baked statically as an o_proj bias. The alpha is tuned to the lowest value that saturates de-refusal.

⚠️ Loading — needs a one-line `glm4_moe` loader patch (or run via `mlx_lm.server`)

mlx-lm's glm4_moe.py hardcodes o_proj to have no bias, so it must be told to build one (backwards-compatible; base models default to False):

# class ModelArgs:  add field
    o_proj_bias: bool = False
# class Attention.__init__:  replace the o_proj line
    self.o_proj = nn.Linear(n_heads * head_dim, dim, bias=getattr(args, "o_proj_bias", False))

The model's config.json sets "o_proj_bias": true. Then load via the patched mlx_lm:

# serve on an OpenAI-compatible endpoint (works with patched mlx-lm)
mlx_lm.server --model gregfrank/GLM-4.5-Air-ULRE-abliterated --port 8080

Point any MCP-capable client (Open WebUI, LibreChat, or LM Studio as an MCP host pointing at the endpoint) at http://127.0.0.1:8080/v1. (LM Studio's bundled MLX engine does not carry this patch, so it won't load the file directly — use mlx_lm.server.)

Notes & caveats

De-refusal validated in both thinking and non-thinking modes (95–96/100 clean), and capability gates were run in thinking mode (math/code preserved or improved). Verified serving via mlx_lm.server.
Research artifact for studying refusal mechanisms / safety-tuning robustness. Use responsibly under the base model's MIT license and applicable law.

Downloads last month: 119

Safetensors

Model size

107B params

Tensor type

BF16

U32

F32

MLX

Hardware compatibility

4-bit

Model tree for gregfrank/GLM-4.5-Air-ULRE-abliterated

Base model

zai-org/GLM-4.5-Air

Quantized

(64)

this model

GLM-4.5-Air-ULRE (abliterated, MLX 4-bit)

Results

Method (ULRE)

⚠️ Loading — needs a one-line glm4_moe loader patch (or run via mlx_lm.server)

Notes & caveats

Model tree for gregfrank/GLM-4.5-Air-ULRE-abliterated

⚠️ Loading — needs a one-line `glm4_moe` loader patch (or run via `mlx_lm.server`)