artindnr/BlueBerry-3
LoRA-style vs openai/gpt-oss-120b. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 36 of 36 layers; norms, biases and embeddings are untouched.
Tensors changed: MLP 0/144, Norms 0/73, Attention 144/324, lm_head 0/1, Embeddings 0/1
2 derivatives checked
- artindnr/BlueBerry-3-GGUF: Quantization
- mradermacher/BlueBerry-3-GGUF: Quantization