trendmicro-ailab/Llama-Primus-Base

Full fine-tune vs meta-llama/Llama-3.1-8B-Instruct. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 0/1, Norms 50/65, MLP 96/96, Attention 128/128, lm_head 1/1

1 derivatives checked