Guilherme34/Darwin-36B-Opus-ABLITERATED-HERETIC

Partial fine-tune vs FINAL-Bench/Darwin-36B-Opus. Partial change: 32% of the weights changed (attention, lm_head, mlp), in 38 of 40 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 0/131, Attention 40/280, MLP 190/280

2 derivatives checked