fixed wrong link in docu (#391)

docs/userguide_surrogates.md

@@ -15,7 +15,7 @@ BoFire offers the following classes of surrogate models.
 [MLP](https://github.com/experimental-design/bofire/blob/main/bofire/surrogates/mlp.py)|a single objective with real-valued inputs|Rich data and black-box function is smooth|Multi layer perceptron
 [MixedSingleTaskGPSurrogate](https://github.com/experimental-design/bofire/blob/main/bofire/surrogates/mixed\_single\_task\_gp.py)|a single objective with categorical and real valued inputs|Limited data and black-box function is smooth|Gaussian process
 [XGBoostSurrogate](https://github.com/experimental-design/bofire/blob/main/bofire/surrogates/xgb.py)|a single objective|Rich data; black-box function does not have to be smooth|xgboost implementation of gradient boosting trees
-[TanimotoGP](https://github.com/experimental-design/bofire/blob/main/bofire/surrogates/tanimoto_gp.py)|a single objective|At least one input feature is a molecule represented as fingerprint|Gaussian process on a molecule space for which Tanimoto similarity determines the similarity between points
+[TanimotoGP](https://github.com/experimental-design/bofire/blob/main/bofire/surrogates/mixed_tanimoto_gp.py)|a single objective|At least one input feature is a molecule represented as fingerprint|Gaussian process on a molecule space for which Tanimoto similarity determines the similarity between points
 
 All of these are single-objective surrogate models. For optimization of multiple objectives at the same time, a suitable [Strategy](https://github.com/experimental-design/bofire/blob/main/bofire/strategies/strategy.py) has to be chosen. Then for each objective a different surrogate model can be specified. By default the SingleTaskGPSurrogate is used.