A collection of notes, usually ML concepts, implementations and papers that I have been reading about recently. Longer then tweets..

Posts

Self Improvement

A comprehensive look at how AI systems are learning to improve themselves
Read More

Deepseek's Janus Models

Multimodal understanding and generation from the same models.
Read More

Modern BERT

A Much-Needed Update to an ML Workhorse
Read More

Gemma Scope

Open Sparse Autoencoders Everywhere All At Once on Gemma 2
Read More

Large Concept Models

LLMs thinking in concepts, not just word by word
Read More