perplexity-ai/pplx-decider-v1.1-27b

Full fine-tune vs Qwen/Qwen3.8-27B. Fine-tune: 95% of the weights changed, across 64 of 64 layers.

Tensors changed: Norms 192/319, Attention 478/556, MLP 271/304, Embeddings 2/4

2 derivatives checked