Fix docstring of FullyBayesianSingleTaskGP (#2894)

Summary:
<!--
Thank you for sending the PR! We appreciate you spending the time to make BoTorch better.

Help us understand your motivation by explaining why you decided to make this change.

You can learn more about contributing to BoTorch here: https://github.com/pytorch/botorch/blob/main/CONTRIBUTING.md
-->

## Motivation

Docstring of `FullyBayesianSingleTaskGP` is wrong (looks like a copy-paste thingy). This PR fixes it.

### Have you read the [Contributing Guidelines on pull requests](https://github.com/pytorch/botorch/blob/main/CONTRIBUTING.md#pull-requests)?

Yes.

Pull Request resolved: #2894

Test Plan: No tests needed, only fix on docstrings.

Reviewed By: saitcakmak

Differential Revision: D77302373

Pulled By: hvarfner

fbshipit-source-id: 3100023293cf38f383523a2c280535527c70e6e9

Author: jduerholt

botorch/models/fully_bayesian.py

@@ -224,18 +224,19 @@ def sample_concentrations(self, **tkwargs: Any) -> tuple[Tensor, Tensor]:
         The prior has a mean value of 1 for each concentration and is very
         concentrated around the mean.
         """
+        d = len(self.indices) if self.indices is not None else self.ard_num_dims
         c0 = pyro.sample(
             "c0",
             pyro.distributions.LogNormal(
-                torch.tensor([0.0] * self.ard_num_dims, **tkwargs),
-                torch.tensor([0.1**0.5] * self.ard_num_dims, **tkwargs),
+                torch.tensor([0.0] * d, **tkwargs),
+                torch.tensor([0.1**0.5] * d, **tkwargs),
             ),
         )
         c1 = pyro.sample(
             "c1",
             pyro.distributions.LogNormal(
-                torch.tensor([0.0] * self.ard_num_dims, **tkwargs),
-                torch.tensor([0.1**0.5] * self.ard_num_dims, **tkwargs),
+                torch.tensor([0.0] * d, **tkwargs),
+                torch.tensor([0.1**0.5] * d, **tkwargs),
             ),
         )
 
@@ -904,13 +905,14 @@ def construct_inputs(
 
 
 class FullyBayesianSingleTaskGP(AbstractFullyBayesianSingleTaskGP):
-    r"""A fully Bayesian single-task GP model with the SAAS prior.
+    r"""A fully Bayesian single-task GP model.
 
     This model assumes that the inputs have been normalized to [0, 1]^d and that
     the output has been standardized to have zero mean and unit variance. You can
     either normalize and standardize the data before constructing the model or use
-    an `input_transform` and `outcome_transform`. A dimension-scaled model
-    [Hvarfner2024vanilla]_ with a Matern-5/2 kernel is used by default.
+    an `input_transform` and `outcome_transform`. A model with a Matern-5/2 kernel
+    and dimension-scaled priors on the hyperparameters from [Hvarfner2024vanilla]_
+    is used by default.
 
     You are expected to use `fit_fully_bayesian_model_nuts` to fit this model as it
     isn't compatible with `fit_gpytorch_mll`.