Passing `is sat` test should expose evaluator if possible

[tnelson]

At the moment, the evaluator is disabled when viewing the (single) instance given when an is sat test passes. I believe that this happens because the solver run is closed after each test to avoid memory bloat. However, not having the evaluator might make it harder to explore the instances shown (e.g., if they are suspected to be incorrect).

The point of this issue is to explore whether or not it's workable to keep the evaluator active (which will lead to more incremental solver states being retained on the back end).