TheCluster/Qwen3.5-27B-Heretic-Text-MLX-mxfp8

Qwen3.5-27B Heretic (Text only)

Quality: quantized (mxfp8, naive, group size: 32)

This is text only abliterated (uncensored) version of Qwen/Qwen3.5-27B

This model supports thinking mode, but it is disabled by default, you can enable thinking via --chat-template-kwargs '{"enable_thinking":true}'

Abliteration metrics

Metric	This model	Original model (Qwen/Qwen3.5-27B)
KL divergence	0.0653	0 (by definition)
Refusals	14/100	94/100

---

Source

This model was converted to MLX format from coder3101/Qwen3.5-27B-heretic using mlx-lm version 0.30.7.