huihui-ai/DeepSeek-R1-Distill-Llama-8B-abliterated
Abliterated vs deepseek-ai/DeepSeek-R1-Distill-Llama-8B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 31 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Attention 31/128, MLP 31/96, Norms 0/65, lm_head 0/1
5 derivatives checked
- cstr/DeepSeek-R1-Distill-Llama-8B-abliterated-Q4_K_M-GGUF: Quantization
- tensorblock/DeepSeek-R1-Distill-Llama-8B-abliterated-GGUF: Quantization
- PocketWeights/DeepSeek-R1-Llama-8B-Abliterated-GGUF: Quantization
- maicog/DeepSeek-R1-Distill-Llama-8B-abliterated-Q4_K_S-GGUF: Quantization
- mergekit-community/DeeperHermes3_R1_D_L3_8b: Full fine-tune