Arjun Kocher
I study deep neural architectures, and their underlying mathematics.
✦
Writing
-
Why Attention Is All You Need
A brief introduction to the seminal Transformer paper.
-
GRPO
A groundbreaking training approach from DeepSeek.
-
Chain of Thought
How reasoning emerges in large language models.
-
Muon
A neural network optimizer.
-
Nested Learning
A new learning paradigm.
-
HOPE
Architecture behind Nested Learning.
Implementations
Notes
