chargoddard/MixtralRPChat-ZLoss

Full fine-tune vs mistralai/Mixtral-8x7B-v0.1. Fine-tune: 99% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 0/1, MLP 768/768, other 32/32, Norms 0/65, Attention 128/128, lm_head 0/1

1 derivatives checked