Mikivis/gpt2-large-lora-sft
LoRA-style vs openai-community/gpt2-large. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of c_attn changed, in 36 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Attention 36/144, Norms 0/146, MLP 0/144, Embeddings 0/2