SicariusSicariiStuff/Impish_LLAMA_4B_Abliterated
Abliterated vs SicariusSicariiStuff/Impish_LLAMA_4B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 20 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 16/96, Attention 20/128, lm_head 0/1
4 derivatives checked
- mradermacher/Impish_LLAMA_4B_Abliterated-i1-GGUF: Quantization
- mradermacher/Impish_LLAMA_4B_Abliterated-GGUF: Same weights, converted
- SicariusSicariiStuff/Impish_LLAMA_4B_Abliterated_GGUF: Quantization
- SicariusSicariiStuff/Impish_LLAMA_4B_Abliterated_GGUF_HA: Quantization
1 other repos hold byte-for-byte identical weight files.