UltimateANJIR-8B-L3-Blackroot-GGUF

This is quantized version of Hastagaras/UltimateANJIR-8B-L3-Blackroot created using llama.cpp

Model Description

So this is a merge of merges. Based on this introverted model: failspy/Llama-3-8B-Instruct-MopeyMule and this extroverted model: Orenguteng/Llama-3-8B-LexiFun-Uncensored-V1

I merged both with my model to add some RP ability, and it turns out the extroverted one is winning.

And also, please don't mind the unusual tokenizer config and the template for all of my models, it's just my preference, you know...for easier testing. And the quantized GGUF might not work with Ollama Chat or Backyard AI... so you have to write the template manually, as it is still the llama 3 instruct template.