usermma/NEXUS-Coder-Abliterated

Partial fine-tune vs fableforge-ai/NEXUS-Coder. Partial change: 45% of the weights changed (attention, embeddings, mlp), in 10 of 28 layers.

Tensors changed: Embeddings 1/1, Norms 0/57, MLP 30/84, Attention 45/196

2 derivatives checked

• mradermacher/NEXUS-Coder-Abliterated-i1-GGUF: Quantization
• mradermacher/NEXUS-Coder-Abliterated-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse