CultriX/Qwen2.5-14B-Qwentangledv2

Partial fine-tune vs arcee-ai/Virtuoso-Small-v2. Partial change: 80% of the weights changed (attention, lm_head, mlp, norms), in 40 of 48 layers.

Tensors changed: Embeddings 0/1, Norms 81/97, MLP 120/144, Attention 280/336, lm_head 1/1

2 derivatives checked