Technical
Technical side projects, often involving live code demos and implementation notes.
-
Exploring plateaus in prior self-play setups.
-
A basic introduction to policy gradient for language models.
-
Results replicating the recent L1 paper.
-
Poking around with gpt-3 and ancient languages
-
A quick go-over of my recent blog changes.
-
I made a fun little animated Ace Attorney AI script generator.
-
Looking into using reinforcement learning to train Pokemon battle bots.
-
Explaining and reproducing some recent work on neural module networks.
-
A toy neural network library.
-
An explanation of the theory and code behind a c++ connect-four AI.
-
How to build and host a very simple chatbot.
-
Looking at generating pokemon with GANs, and hosting the resulting models.