diff --git a/docs/modules/validate.rst b/docs/modules/validate.rst
index 60a466a..6bec348 100644
--- a/docs/modules/validate.rst
+++ b/docs/modules/validate.rst
@@ -15,10 +15,29 @@ Classes
    :template: class_with_private.rst
 
     DatasetSplitter
+
+Scoring
+-------
+.. currentmodule:: tpcp.validate
+
+
+.. autosummary::
+   :toctree: generated/validate
+   :template: class_with_private.rst
+
     Scorer
     Aggregator
-    MeanAggregator
-    NoAgg
+    FloatAggregator
+
+.. currentmodule:: tpcp.validate
+
+.. autosummary::
+   :toctree: generated/validate
+   :template: function.rst
+
+    mean_agg
+    no_agg
+
 
 Functions
 ---------
diff --git a/tpcp/validate/_scorer.py b/tpcp/validate/_scorer.py
index dc36bcd..cfdd8e1 100644
--- a/tpcp/validate/_scorer.py
+++ b/tpcp/validate/_scorer.py
@@ -158,10 +158,6 @@ def aggregate(self, /, values: Sequence[float], datapoints: Sequence[Dataset]) -
         return float(vals)
 
 
-mean_agg = FloatAggregator(np.mean)
-mean_agg.__doc__ = """Aggregator that calculates the mean of the values."""
-
-
 class _NoAgg(Aggregator[Any]):
     """Wrapper to wrap one or multiple output values of a scorer to prevent aggregation of these values.
 
@@ -175,7 +171,7 @@ class _NoAgg(Aggregator[Any]):
     --------
     >>> def score_func(pipe, dataset):
     ...     ...
-    ...     return {"score_val_1": score, "some_metadata": NoAgg(metadata)}
+    ...     return {"score_val_1": score, "some_metadata": no_agg(metadata)}
     >>> my_scorer = Scorer(score_func)
 
     """
@@ -185,9 +181,23 @@ def aggregate(self, /, values: Sequence[Any], datapoints: Sequence[Dataset]) ->
         return NOTHING
 
 
-no_agg = _NoAgg()
+# We wrap the existing aggregators in functions, so that we can properly document them.
+_no_agg = _NoAgg()
+
+
+def no_agg(value: Any) -> _NoAgg:
+    return _no_agg(value)
+
+
 no_agg.__doc__ = _NoAgg.__doc__
 
+_mean_agg = FloatAggregator(np.mean)
+
+
+def mean_agg(value: float) -> FloatAggregator:
+    """Calculate the mean of the values."""
+    return _mean_agg(value)
+
 
 class Scorer(Generic[PipelineT, DatasetT, T], BaseTpcpObject):
     """A scorer to score multiple data points of a dataset and average the results.