Anserwise/AWAXIS-Think-31B

Partial fine-tune vs TeichAI/gemma-4-31B-it-Claude-Opus-Distill-v2. Partial change: 67% of the weights changed (mlp, norms), in 60 of 60 layers.

Tensors changed: Embeddings 0/4, Norms 120/523, other 0/62, MLP 180/261, Attention 0/338

2 derivatives checked

• mradermacher/AWAXIS-Think-31B-i1-GGUF: Quantization (fewer layers)
• mradermacher/AWAXIS-Think-31B-GGUF: Quantization (fewer layers)

On the Hub · Downloads on Model Pulse