exalandru/GPT-OSS-Coder-GGUF
LoRA-style vs openai/gpt-oss-120b. exalandru/GPT-OSS-Coder-GGUF is a GGUF (MXFP4/BF16) of different weights. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, router, v_proj changed, in 8 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/73, MLP 8/72, Attention 32/288, lm_head 0/1, Embeddings 0/1