Huihui-Qwen3.5-9B-abliterated-mlx-8bit

MLX-VLM 8bit export of huihui-ai/Huihui-Qwen3.5-9B-abliterated for Apple Silicon workflows, including LM Studio and local mlx_vlm usage.

Overview

• Variant: 8bit
• Repository payload at upload time: 10.5G
• Repository file count: 13
• Effective quantization observed during conversion: 8.864 bits per weight
• Format: mlx-vlm model package

Compatibility

• Uses the corrected Qwen VL chat template with image token placeholders.
• Uses <|im_end|>-compatible stop token settings for cleaner chat termination in MLX/LM Studio.

Validation

• Local text generation smoke test: passed
• Local image-input smoke test: passed
• Local black-box abliterated check: 6/6 non-refused
• Refusal rate: 0.0
• Actionable non-refused cases: 6
• Median cleaned response length: 1211 chars

Behavior Notes

• This variant preserved the abliterated behavior on the local 6-case regression set used during conversion validation.
• These checks are behavioral acceptance tests, not a formal guarantee of identical outputs to the source checkpoint.

Usage

mlx_vlm.generate \
  --model /path/to/Huihui-Qwen3.5-9B-abliterated-mlx-8bit \
  --prompt "你好" \
  --max-tokens 256

For behavior-focused checks, it is safer to disable thinking output so refusal scoring is based on the final answer instead of the thinking trace.

mlx_vlm.generate \
  --model /path/to/Huihui-Qwen3.5-9B-abliterated-mlx-8bit \
  --prompt "你好" \
  --max-tokens 256 \
  --processor-kwargs '{"enable_thinking": false}'

LM Studio

• If LM Studio has an older cached copy, refresh or re-download the repository so the latest chat template and config are picked up.
• These repositories are meant for mlx-vlm / Apple MLX runtimes rather than Transformers CPU inference.