netcat420/MFANNv0.21.10

Partial fine-tune vs mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated. Partial change: 45% of the weights changed (attention, mlp, norms), in 32 of 32 layers.

Tensors changed: Embeddings 0/1, Norms 9/65, MLP 38/96, Attention 128/128, lm_head 0/1

2 derivatives checked

• featherless-ai-quants/netcat420-MFANNv0.21.10-GGUF: Quantization
• mradermacher/MFANNv0.21.10-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse