dots3-note shows test-time learning: explores, self-corrects, reuses knowledge in unseen environments

SarahAnnabels · x · 2026-08-30

Dropped into an unfamiliar environment, dots3-note Preview can explore through interaction, form and test hypotheses, update its own memory, correct stale assumptions, and reuse what it learns. The same behavior reportedly generalized to Slay the Spire 2, an environment it was not trained on—pointing to genuine test-time learning rather than pretrained knowledge alone.

Related event: dots3-note's Test-Time Learning and TEMPO Recursive Self-Evaluation(2 posts)→

Original post →

More from Models

Models channel →