Avesed/Qwopus3.6-27B-v2-abliterated

Qwopus3.6-27B-v2-abliterated

Refusal-ablated ("abliterated") build of Jackrong/Qwopus3.6-27B-v2, a Qwen3.5 hybrid (GatedDeltaNet linear-attention + gated full-attention) VL reasoning model.

Method

Refusal-direction orthogonalization, no fine-tuning: the refusal direction is estimated from harmful/harmless prompt activations and orthogonalized out of the residual-stream write matrices (o_proj / down_proj) at layer 26. Vision tower untouched.

• —Refusal rate: 100% -> 8%
• —General capability: preserved (evals below)

Evaluation

bf16 weights. For a ~26 GB INT4 vLLM-deployable build see Qwopus3.6-27B-v2-abliterated-int4. And GGUF Qwopus3.6-27B-v2-abliterated-GGUF.

MTP head

The Multi-Token-Prediction (mtp) head is included (for speculative decoding). Its residual-write matrices (selfattn.oproj, mlp.down_proj) are abliterated with the same refusal direction as the main layers.