Apply new optimal set of settings after tuning

[amidgol]

This change is based on the results of the tuning paper. We found a new set of configs to be optimal, detecting a higher number of faults while keeping a competitive degree of line and branch coverage. This setting is referred to as "best faults" in the paper.

[arcuri82]

@amidgol some E2E tests fail after the change of parameters. This shouldn't really happen (but might be issue with the E2E). Few seem related to AdaptiveHypermutation, so you might need to have a chat or help from @man-zhang for looking into it