Null-Guard/LFM2.5-230M-Uncensored-GGUF
Abliterated vs LiquidAI/LFM2.5-230M. Null-Guard/LFM2.5-230M-Uncensored-GGUF is a GGUF (F16) of different weights. Looks like abliteration (refusal removal): only the output projections (embed_tokens, out_proj, w2) changed, in 14 of 14 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 1/2, other 0/16, MLP 14/42, Norms 0/40, Attention 6/24