Mechanistic interpretability researchers applying causality theory to LLMs

30papers.com – Ilya's 30 essential ML papers, in a beginner friendly format

DeepSeek open-sources inference optimizations with 60–85% faster generation [pdf]

CAPTCHAs can still detect AI agents

There Will Be a Scientific Theory of Deep Learning

Talos: Hardware accelerator for deep convolutional neural networks

DjVu and its connection to Deep Learning (2023)