Qwen3.8-27B Abliterated MLX MXFP4

Vision-enabled MLX conversion of Qwen/Qwen3.8-27B, pinned to revision 1d4bf0f2ff6012fd82039f2fa52739d0dd7c60c0, with a measured refusal-direction edit.

  • Language model: MXFP4, group size 32
  • Vision tower: unchanged same-revision BF16 weights
  • Edit: completion-projected global direction, strength 2.5, norm-preserving, 128 modules
  • Runtime: mlx-vlm 0.6.3 or newer

On a frozen 12-case held-out suite, measured refusal markers changed from 12/12 in the exact quantized parent to 0/12, while benign-sensitive refusal stayed 0/12 and utility stayed 6/6. This is scoped evidence, not a claim that the model is universally โ€œuncensored.โ€

The Hub's approximately 5.5B safetensors count reflects packed MXFP storage elements; the underlying architecture remains the full 27B model.

python -m mlx_vlm.generate \
  --model Shiftedx/Qwen3.8-27B-Abliterated-MLX-MXFP4 \
  --image image.jpg --prompt "Describe this image."

This edit reduces refusal behavior and may increase harmful or incorrect outputs. Treat prompts, images, and outputs as untrusted; do not submit secrets, and sandbox tools or generated code with least-privilege access. This conversion adds no telemetry or remote execution.

The upstream Apache-2.0 license and model limitations continue to apply.

Downloads last month
464
Safetensors
Model size
27B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for Shiftedx/Qwen3.8-27B-Abliterated-MLX-MXFP4

Base model

Qwen/Qwen3.8-27B
Quantized
(1,826)
this model

Collection including Shiftedx/Qwen3.8-27B-Abliterated-MLX-MXFP4