inclusionAI/ZwZ-8B
Partial fine-tune vs Qwen/Qwen3-VL-8B-Instruct. Partial change: 86% of the weights changed (attention, embeddings, mlp, norms), in 36 of 36 layers.
Tensors changed: Embeddings 3/4, Norms 172/261, MLP 213/232, Attention 252/252, lm_head 0/1
5 derivatives checked
- mradermacher/ZwZ-8B-i1-GGUF: Quantization (fewer layers)
- mradermacher/ZwZ-8B-GGUF: Same weights, converted (fewer layers)
- noctrex/ZwZ-8B-GGUF: Same weights, converted (fewer layers)
- MuXodious/ZwZ-8B-PaperWitch-heresy: Abliterated
- megabytes/ZwZ-8B-heretic: Abliterated