lm_understand-logs

forest

FOREST is a research agent for long-horizon science, where evidence gets lost, errors compound, and goals drift. It explores ideas, reads literature, forms hypotheses, runs real experiments, analyzes results, revises paths, and builds a full paper—all in an editable graph you can inspect, branch, redirect, with claims and figures linked to evidence

memoir

Weights that never freeze. A continual-learning architecture, and the pre-registered experiment that falsified its central claim. https://arxiv.org/abs/2607.20792

onecanvas

[NeurIPS 2026 Spotlight] Official code and checkpoints for OneCanvas: 3D Scene Understanding via Panoramic Reprojection.

pointdit