jnjn/Ornith-1.5-35B-A3B-Spark-Abliterated-VL-MTP-NVFP4
Partial fine-tune vs ornith-ai/Ornith-1.5-35B-A3B-NVFP4. Partial change: 31% of the weights changed (attention, lm_head, mlp), in 40 of 40 layers.
Tensors changed: Attention 298/652, MLP 631/2092, Embeddings 0/5, Norms 0/247, lm_head 1/3, other 0/1