Stange features in light jet rejection rates of MV2c10 in rel21

Multi-b-jet and Di-b-jet resonance analysis teams have observed strange features in the light jet rejection rates of MV2c10 taggers. Please see attached slides. Even though from an analyzer's point of view we can just switch to DL1, it is useful to understand what is going on regarding the MV2c10 taggers.

We will update studies/results in the Jira ticket and any suggestions or comments are really appreciated!