CodeGoat24/WorldReward-qwen35-9b
Partial fine-tune vs Qwen/Qwen3.5-9B. Partial change: 84% of the weights changed (attention, lm_head, mlp, norms), in 32 of 32 layers.
Tensors changed: lm_head 1/1, Embeddings 0/4, MLP 96/208, Attention 183/332, Norms 72/215
3 derivatives checked
- mradermacher/WorldReward-qwen35-9b-i1-GGUF: Quantization
- mradermacher/WorldReward-qwen35-9b-GGUF: Same weights, converted
- prithivMLmods/WorldReward-qwen35-9b-GGUF: Same weights, converted