Securelayer7/Ling-3.0-tiny-Uncensored-Abliterated
Abliterated vs inclusionAI/Ling-3.0-tiny. Looks like abliteration (refusal removal): only the output projections (dense, down_proj, o_proj) changed, in 22 of 24 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Attention 22/252, Norms 0/79, MLP 502/2667, Embeddings 0/1, lm_head 0/1
2 derivatives checked
- mradermacher/Ling-3.0-tiny-Uncensored-Abliterated-i1-GGUF: Not comparable yet
- mradermacher/Ling-3.0-tiny-Uncensored-Abliterated-GGUF: Not comparable yet