Load a fixture JSON to visualize GP posterior slices
About this plot
Each subplot shows the GP posterior prediction (mean ± confidence band) as a single parameter varies, with all other parameters held fixed. Parameters are sorted by feature importance (shorter kernel lengthscale = more important).
The blue curve is the posterior mean. The shaded band is the 95% CI. Training points are shown as dots — hover to highlight nearest neighbors across subplots.
Interactivity:
Outcome selector — switch between objectives
Sliders — adjust "held fixed" values for other parameters
Pivoting — click the mean curve to set that parameter's value in all other subplots, exploring the response surface from that operating point. Dot opacity reflects kernel correlation (more distant = more transparent).
Training dots — click to snap all sliders to that trial's parameter values