mergekit-community/MN-Hekate-Limenoskopos-17B

Partial fine-tune vs mergekit-community/MN-Hekate-Damnomeneia-17B. Partial change: 53% of the weights changed (attention, mlp, norms), in 32 of 56 layers.

Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 17/113, MLP 96/168, Attention 128/224

2 derivatives checked