longtermrisk/Qwen3-8B-german-city-names-second-third-v2-sft-seed3-epoch3
LoRA-style vs unsloth/Qwen3-8B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 12 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 36/108, Attention 48/144, lm_head 0/1