vanta-research/atom-olmo3-7b
LoRA-style vs allenai/Olmo-3-7B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj changed, in 1 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, MLP 0/96, Norms 0/129, Attention 1/128, lm_head 0/1
2 derivatives checked
- mradermacher/atom-olmo3-7b-i1-GGUF: Quantization
- mradermacher/atom-olmo3-7b-GGUF: Same weights, converted