nkpz/T3Q-qwen2.5-14b-v1.0-e3-Uncensored-DeLMAT
LoRA-style vs JungZoona/T3Q-qwen2.5-14b-v1.0-e3. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of gate_proj changed, in 47 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/97, MLP 47/144, Attention 0/336, lm_head 0/1