coder3101/Qwen3-VL-4B-Instruct-heretic
Abliterated vs Qwen/Qwen3-VL-4B-Instruct. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 34 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/4, Norms 0/249, MLP 27/220, Attention 34/240
1 derivatives checked
- matrixportalx/Qwen3-VL-4B-Instruct-heretic-Q4_K_M-GGUF: Quantization (fewer layers)