Trajectory Geometry of Transformer Representations Across Layers

Prosthetic cognition architecture for AI agents. Deterministic scaffolding over probabilistic reasoning.

Mechanistic interpretability of multilingual reasoning in transformers. 170+ causal intervention experiments across 4 model families.

Code for 'Reliable Measures of Spread in High Dimensional Latent Spaces' (ICML 2023)

A 2-D map of GPT-2's token embeddings you can poke at. Pick two axes, project the vocabulary, and see where analogies, cyclic features, and hubness show up (and where the 2-D view is lying to you). Companion to a writeup on which "cyclic" concepts actually form circles.

Mechanistic interpretability of transformer hallucinations via attention flow, residual stream geometry, and head-level attribution analysis.

Personal learning notebook on latent space engineering, representation geometry, and exploratory research notes.

Code for 'Exploring the Impact of a Transformer's Latent Space Geometry on Downstream Task Performance' (arXiv:2406.12159)

Visualizing Modern LLM Mechanics, Loss Landscapes & HPC Topologies

Com la tokenització fractura la morfologia catalana i si una segmentació conscient dels morfemes recupera la geometria. Provat en 3 llengües indoeuropees (català, castellà, anglès): el català es fragmenta ~1,7× més que l'anglès; forçar el tall morfèmic recupera la composicionalitat (robust a portadora i replicat en castellà).