Blog

Essays and technical notes on research, engineering, and lessons from the lab.

Using Logit Space of VLMs for Attention to Detail

RESEARCH·2026 · 03 · 18·11 MIN

Notes from a DARPA Triage Challenge perception pipeline: why attention maps alone can't tell you what a VLM actually saw, and how projecting intermediate hidden states through the frozen LM head — the logit lens — gets you a classifier that reasons over trauma cues instead of whole images.

continue reading →

Why JEPA Beat ImageNet Pretraining for Spacecraft Pose

RESEARCH·2026 · 03 · 12·13 MIN

Notes from SPACE-HOP: how embedding-predictive pretraining and a Hopf-fibration rotation grid reshaped our 6-DoF pose estimation pipeline for uncooperative spacecraft.

continue reading →