SC117/LFM2.5-2.6B-Uncensored
Abliterated vs LiquidAI/LFM2.5-2.6B. Looks like abliteration (refusal removal): only the output projections (out_proj, w2) changed, in 14 of 30 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/2, other 4/66, MLP 11/90, Norms 0/76, Attention 6/32
4 derivatives checked
- mradermacher/LFM2.5-2.6B-Uncensored-i1-GGUF: Quantization
- mradermacher/LFM2.5-2.6B-Uncensored-GGUF: Same weights, converted
- amarsadan/AMARSADN-AI_V1: LoRA-style
- amarsadan/X-YEMEN_HACK-AI_V1.GGUF: Adapter
1 other repos hold byte-for-byte identical weight files.