SicariusSicariiStuff/Llama-3.3-8B-Instruct-128K_Abliterated
Abliterated vs shb777/Llama-3.3-8B-Instruct-128K. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 18 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 17/96, Attention 18/128, lm_head 0/1
5 derivatives checked
- bartowski/SicariusSicariiStuff_Llama-3.3-8B-Instruct-128K_Abliterated-GGUF: Same weights, converted
- mradermacher/Llama-3.3-8B-Instruct-128K_Abliterated-i1-GGUF: Quantization
- mradermacher/Llama-3.3-8B-Instruct-128K_Abliterated-GGUF: Same weights, converted
- SicariusSicariiStuff/Llama-3.3-8B-Instruct-128K_Abliterated_GGUF: Quantization
- blascotobasco/Llama-3.3-Octet: LoRA-style
1 other repos hold byte-for-byte identical weight files.