Goekdeniz-Guelmez/Josiefied-Qwen3-0.6B-gabliterated-v1
Abliterated vs Goekdeniz-Guelmez/Qwen3-0.6B-gabliterated. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 26 of 28 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/113, MLP 26/84, Attention 26/112
2 derivatives checked
- mradermacher/Josiefied-Qwen3-0.6B-gabliterated-v1-i1-GGUF: Quantization
- mradermacher/Josiefied-Qwen3-0.6B-gabliterated-v1-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.