Babsie/Trickster-Theta-4-70B
Full fine-tune vs NousResearch/Hermes-2-Theta-Llama-3-70B. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Norms 161/161, MLP 240/240, Attention 320/320
4 derivatives checked
- mradermacher/Trickster-Theta-4-70B-i1-GGUF: Quantization
- Babsie/Trickster-Theta-4-70B-GGUF: Same weights, converted
- mradermacher/Trickster-Theta-4-70B-GGUF: Quantization
- Babsie/Trickser_Theta_4_70B_calibrated_gguf: Quantization
1 other repos hold byte-for-byte identical weight files.