nkpz/L3.1-Siithamo-v0.3-8B-Uncensored-DeLMAT
LoRA-style vs kromvault/L3.1-Siithamo-v0.3-8B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/65, MLP 64/96, Attention 0/128
2 derivatives checked
- mradermacher/L3.1-Siithamo-v0.3-8B-Uncensored-DeLMAT-i1-GGUF: Quantization
- mradermacher/L3.1-Siithamo-v0.3-8B-Uncensored-DeLMAT-GGUF: Same weights, converted