Deep Learning Beyond the Plateau | Substack
An interpretive blog on deep learning research, focusing on why models are designed the way they are and how structure shapes learning behavior.
Visit Deep Learning Beyond the Plateau | Substack
Recent posts
- Teaching Models to Behave: A Field Guide to LLM Post-Training Techniques
- GIST: How to Pick a Dataset That's Both Good and Diverse
- CUDA Made Simple: A Short, Practical GPU Programming Guide
- RWKV: Reinventing RNNs for the Transformer Era
- Infini-Gram: The Return of the Simplest Language Model
- Engram: The Other Kind Of Sparsity Your LLM Is Missing
- The Neural Tangent Kernel: What It Tells Us About Neural Networks (And What It Doesn't)
- Nemotron: Building Production-Ready Hybrid Models
- Mamba: Linear-Time Sequence Modeling with Selective State Spaces
- Recursive Reasoning with Tiny Networks