internlm/Intern-Decision-4B
Partial fine-tune vs Qwen/Qwen3.5-4B. Partial change: 79% of the weights changed (attention, mlp, norms), in 32 of 32 layers.
Tensors changed: Embeddings 0/4, MLP 96/196, Attention 201/320, Norms 76/203
5 derivatives checked
- bombdefuser-124/Intern-Decision-4B-GGUF: Same weights, converted
- yeaay/Intern-Decision-4B-GGUF: Same weights, converted
- mertkayacs/Wahler-4B: LoRA-style
- mertkayacs/Karar-4B: LoRA-style
- mertkayacs/Deem-4B: LoRA-style