Curious what you might consider “adequate shrinking”?

Horshoe priors, partial pooling, something more?

I realize that might be highly subject

I guess this depends on the problem at hand.

But I was thinking about a typical hierarchical model with partial pooling and standard weakly informative priors.