Interpreting glmnet plots

Here's the tuning plot for the custom tuned glmnet model you created in the last exercise. For the overfit dataset, which value of alpha is better?