Sao10K

/

L3-8B-Lunaris-v1

like 79

..Title..

• Handles everything well
• Better Reasoning Quality than a 13B model or a (2.5 bpw) elx2 70B (llama 3) model
• Handles Multiple Character (speaks as npc when needed)
• No format Issue (unlike the Fimbulvetr v2)
• A lot better (hard to consider that it's a 8B model)
• Follows Character Card very well
• it's almost like a Human (kinda poetic sometimes)