August 5, 2025
The Challenges of Superposition in LLMs
How neural networks represent more features than they have neurons, and why this complicates interpretability efforts.
Research & Writings
August 5, 2025
Inductive Heads in Transformer Models
What are inductive heads and in-context learning.
August 4, 2025
A Comprehensive Architectural Review of Large Language Models
This report provides a comprehensive, expert-level deconstruction of this architectural evolution.