Naphula/Human-Like-LLama3-8B-Instruct-MPOA
Abliterated vs HumanLLMs/Human-Like-LLama3-8B-Instruct. 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: Embeddings 0/1, Norms 0/65, MLP 32/96, Attention 32/128, lm_head 0/1
2 derivatives checked
- mradermacher/Human-Like-LLama3-8B-Instruct-MPOA-i1-GGUF: Quantization
- mradermacher/Human-Like-LLama3-8B-Instruct-MPOA-GGUF: Same weights, converted