byroneverson/internlm2_5-20b-chat-abliterated

Partial fine-tune vs internlm/internlm2_5-20b-chat. Partial change: 36% of the weights changed (attention, embeddings, mlp), in 48 of 48 layers.

Tensors changed: Attention 48/96, Norms 0/97, MLP 48/144, Embeddings 1/1, lm_head 0/1

2 derivatives checked