VerusCommunity/llama-3-verus-revision-2
039
Changes:
Smarter training data
Trained to understand the mission of Verus Trained to understand the current state of Verus's technology better
Better LMStudio Config (Use the plain_turns config in THIS repo for this model)
Less-cluttered Repo

Known issue: model will go braindead if it sees chatml tokens. Do not use those. The next version will use a more minimalistic format since chatml apparently degrades performance so severely.
