MultiverseComputingCAI/HyperNova-60B
Partial fine-tune vs openai/gpt-oss-120b. Partial change: 35% of the weights changed (attention, lm_head, norms), in 5 of 32 layers.
Tensors changed: Norms 10/65, Attention 40/288, lm_head 1/1, Embeddings 0/1
5 derivatives checked
- MultiverseComputingCAI/Hypernova-60B-2602: Full fine-tune
- MultiverseComputingCAI/Hypernova-60B-2605: Full fine-tune
- mradermacher/HyperNova-60B-i1-GGUF: Quantization
- mradermacher/HyperNova-60B-GGUF: Quantization
- blascotobasco/HyperNova-64E: Same weights, repacked