Neurotype
Evaluation dashboard

Is it getting better, and why?

Metric: character error rate (CER), the edit distance between the output and the ground truth divided by the length of the ground truth. Lower is better. 0 means a perfect reconstruction.

Layer 1 · What

every version, every run

Each dot is one full run. The line connects the mean of each version. Dots of the same version differ only because the model is not perfectly repeatable.

Layer 2 · Why

Prompt for this run