Which hyper parameter to select after nested cross validation?

Question

I am currently doing a nested cross-validation for lasso regression to determine the best lambda value.

I am using the mlr3 package with the documentation from https://mlr3book.mlr-org.com/optimization.html#sec-nested-resampling

In the example, the hyperparameters selected and tested on the outer fold were:

iteration   cost      gamma classif.ce

      1 -11.512925 -11.512925  0.4567227
      1 -11.512925   5.756463  0.4567227
      1  -5.756463 -11.512925  0.4567227
      1   0.000000   5.756463  0.4567227
      1   5.756463 -11.512925  0.2747899
     ...                                  
      3   0.000000 -11.512925  0.4678571
      3   0.000000  -5.756463  0.2151261
      3   0.000000   5.756463  0.4678571
      3   5.756463   0.000000  0.4678571
      3  11.512925  -5.756463  0.1941176

So how do I determine the best hyperparameters (cost and gamma in this case) for my final model?