-
Enna is now using two separate LLMs at the same time. one for chat, and one for taking notes, function calling and intent detection. that way I can have the main chat model unencumbered by the instruct plague, and use a high gptslop signal benchmark destroyer for instruct tasks!On twitter.com
♻️ 2 Retweets
❤️ 50 Favorites
Mood -3 🙁