My methods, side by side with simple baselines.

Each card is one experiment from my public repositories: what was tested, under which conditions, how it compares with a reference (with its uncertainty), and where it stands. Some results reproduced, in some a simpler baseline did better, and some are still open. They all stay on the page, and every number comes from a committed file at the commit shown.

Research software. Synthetic data is labelled synthetic. No claims of quantum error correction, quantum advantage, hardware performance or consciousness.