I’ve created this page to host some deep dives into complex systems, starting with a walkthrough of the internal parameter ‘world’ of the Llama 2 large language model. My background is in spectroscopy and quantum mechanical modeling/analysis of advanced materials. This means that I like to pull apart and visualize high dimensional objects to learn about how they work - and then to use similar pieces to build something new.
Posts
-
The Chatbot Architecture of Tomorrow – (3) What Are the Solutions?
-
The Chatbot Architecture of Tomorrow – (2) Limits of Chatbot Cognition
-
The Chatbot Architecture of Tomorrow – (1) Setting the Stage
-
Are two dummies better than one?
-
Llama surprises
-
Visualizing the inner world of a large language model (Llama 2 7B)
subscribe via RSS