Back to AIBriefs
AnalysisAI Models

Qwable 27B fine-tune replicates Fable 5 reasoning locally

Qwable 27B is a full fine-tune of Alibaba's Qwen3.6-27B on a Fable 5-style reasoning dataset. An abliterated version removes refusal behavior by surgically modifying weights.

·
Jun 23, 10:01 PM