AveLabs/Qwen3.8-27B-Perplexing-Pegasus

Partial fine-tune vs Qwen/Qwen3.8-27B. Partial change: 89% of the weights changed (attention, embeddings, mlp, norms), in 64 of 64 layers.

Tensors changed: Norms 180/325, Attention 465/560, MLP 272/307, Embeddings 2/5, lm_head 0/1, other 0/1

1 derivatives checked

• AveLabs/Qwen3.8-27B-Perplexing-Pegasus-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse