0xBakeer/TandemLLM-Qwen3.8-27B-DFlash2-b16
Full fine-tune vs z-lab/Qwen3.8-27B-DFlash2. Fine-tune: 93% of the weights changed, across 5 of 5 layers.
Tensors changed: other 1/4, Norms 14/22, Attention 30/30, MLP 25/25
Full fine-tune vs z-lab/Qwen3.8-27B-DFlash2. Fine-tune: 93% of the weights changed, across 5 of 5 layers.
Tensors changed: other 1/4, Norms 14/22, Attention 30/30, MLP 25/25