mlx-community/gpt2-base-mlx
LoRA-style vs openai-community/gpt2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of c_proj changed, in 12 of 12 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/50, Attention 12/36, MLP 0/24, Embeddings 0/2