OccultAI/Musecuilo-12B-Model_Stock

Partial fine-tune vs LatitudeGames/Muse-12B. Partial change: 89% of the weights changed (attention, mlp, norms), in 40 of 40 layers.

Tensors changed: Embeddings 0/1, Norms 4/81, MLP 120/120, Attention 160/160, lm_head 0/1

2 derivatives checked