Increased accuracy for frequency nodelocks in AI solver

When nodelocking frequencies in AI custom solver the accuracy is limited to whole percentage points. Could you increase that to one decimal point?

Example use case:

SRP BTN vs BB

I want to study c-betting on a series of flops where my opponent model check-raises 10% too much (not in absolute terms, but relative to his GTO range).

On one board his GTO XR% is 5.7%. I want to lock him to (relative) 10% more than that, so 5.7% + 0.57% = 6.3%. But right now I can only drag the slider to 6%, which is too crude.

On another board his XR% is 15%. Now I want to lock at 15% + 1.5% = 16.5%, but the closest I can get is 16% or 17%.

Such work would be more accurate if we could have a decimal on the frequency locks, which is similar to how frequencies are reported throughout the app.

Thanks in advance!