[Bug] I-BNN tutorial benchmarks need updating

[slishak-PX]

🐛 Bug

In the documentation for BoTorch 0.11.3, I-BNN is shown to outperform standard GPs with Matern and RBF kernels:
https://botorch.org/v/0.11.3/tutorials/ibnn_bo

In the 0.12.0 documentation, the Matern kernel's performance increases substantially and dominates the others, contradicting the text:
https://botorch.org/v/0.12.0/tutorials/ibnn_bo

I initially assumed this was due to the dimension-dependent LogNormal prior, but it doesn't look like this is used in the tutorial.

[saitcakmak]

Thanks for flagging. Looks like it was switch to LogEI #2499 that helped Matern kernel. I generally dislike such comparisons in tutorials. We need tutorials to be runnable with little compute for the CI to run within a reasonable time, and comparing multiple methods on a single replication just produces unreliable results.