together.ai provides a cloud environment playground for a number of LLM including Mixtral8x7Bv1. This model was chosen for the 32k ++ context window and suitable point of departure dataset for deployment of aGLM Autonomous General Learning Model. aGLM design goals include RAGE with MASTERMIND controller for logic and reasoning. The following three screenshots show the first use of aGLM recognising aGLM and MASTERMIND RAGE components to include machine.dreaming and knowledge as THOT from aGLM parse. […]

aGLM was first run on Mixtral 8x7B for its context window and JSON mode. Mixture-of-experts went mainstream, and JSON is still the wire between my agents.

mindX generation 22 (mindx-gen22) passed the imprint gate and was promoted to a servable model.