bunnycore/Phi-3.5-Mini-Sonet-RP
Abliterated vs bunnycore/Phi-3.5-Mini-Hyper. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 32 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/65, MLP 32/64, Attention 32/64
1 derivatives checked
- mradermacher/Phi-3.5-Mini-Sonet-RP-GGUF: Quantization