diff --git a/optax/_src/lookahead.py b/optax/_src/lookahead.py
index 98d6103b5..5ef299424 100644
--- a/optax/_src/lookahead.py
+++ b/optax/_src/lookahead.py
@@ -30,8 +30,8 @@ class LookaheadState(NamedTuple):
 
   Attributes:
     fast_state (:class:`optax.OptState`): Optimizer state of the fast optimizer.
-    steps_since_sync (``Union[jax.Array, int]``): Number of fast optimizer steps taken since slow and fast
-      parameters were synchronized.
+    steps_since_sync (``Union[jax.Array, int]``): Number of fast optimizer steps
+      taken since slow and fast parameters were synchronized.
   """
   fast_state: base.OptState
   steps_since_sync: Union[jax.Array, int]
diff --git a/optax/_src/wrappers.py b/optax/_src/wrappers.py
index b901a5228..28d0cd148 100644
--- a/optax/_src/wrappers.py
+++ b/optax/_src/wrappers.py
@@ -90,18 +90,17 @@ class ApplyIfFiniteState(NamedTuple):
   """State of the `GradientTransformation` returned by `apply_if_finite`.
 
   Attributes:
-    notfinite_count (``Union[jax.Array, int]``): Number of consecutive gradient updates containing an Inf or
-      a NaN. This number is reset to 0 whenever a gradient update without an Inf
-      or a NaN is done.
-    last_finite (``Union[jax.Array, int]``): Whether or not the last gradient update contained an Inf or a
-      NaN.
-    total_notfinite (``Union[jax.Array, int]``): Total number of gradient updates containing an Inf or
-      a NaN since this optimizer was initialised. This number is never reset.
-    inner_state (:class:`optax.OptState`): The state of the inner `GradientTransformation`.
+    notfinite_count (``Union[jax.Array, int]``): Number of consecutive gradient
+      updates containing an Inf or a NaN. This number is reset to 0 whenever a
+      gradient update without an Inf or a NaN is done.
+    last_finite (``Union[jax.Array, int]``): Whether or not the last gradient
+      update contained an Inf or a NaN.
+    total_notfinite (``Union[jax.Array, int]``): Total number of gradient
+      updates containing an Inf or a NaN since this optimizer was initialised.
+      This number is never reset.
+    inner_state (:class:`optax.OptState`): The state of the inner 
+      `GradientTransformation`.
   """
-  # TODO(optax-dev): notfinite_count, last_finite and inner_state used to be
-  # annotated as `jnp.array` but that is not a valid annotation (it's a function
-  # and secretely resolved to `Any`. We should add back typing.
   notfinite_count: Union[jax.Array, int]
   last_finite: Union[jax.Array, int]
   total_notfinite: Union[jax.Array, int]
@@ -176,16 +175,17 @@ class MultiStepsState(NamedTuple):
   """State of the `GradientTransformation` returned by `MultiSteps`.
 
   Attributes:
-    mini_step (``Union[jax.Array, int]``): current mini-step counter. At an update, this either increases by
-      1 or is reset to 0.
-    gradient_step (``Union[jax.Array, int]``): gradient step counter. This only increases after enough
-      mini-steps have been accumulated.
-    inner_opt_state (:class:`optax.OptState`): the state of the wrapped optimiser.
+    mini_step (``Union[jax.Array, int]``): current mini-step counter. At an
+      update, this either increases by 1 or is reset to 0.
+    gradient_step (``Union[jax.Array, int]``): gradient step counter. This only
+      increases after enough mini-steps have been accumulated.
+    inner_opt_state (:class:`optax.OptState`): the state of the wrapped
+      optimiser.
     acc_grads (``jax.Array``): accumulated gradients over multiple mini-steps.
-    skip_state (``chex.ArrayTree``): an arbitrarily nested tree of arrays. This is only
-      relevant when passing a `should_skip_update_fn` to `MultiSteps`. This
-      structure will then contain values for debugging and or monitoring. The
-      actual structure will vary depending on the choice of
+    skip_state (``chex.ArrayTree``): an arbitrarily nested tree of arrays. This
+      is only relevant when passing a `should_skip_update_fn` to `MultiSteps`.
+      This structure will then contain values for debugging and or monitoring.
+      The actual structure will vary depending on the choice of
       `ShouldSkipUpdateFunction`.
   """
   mini_step: Union[jax.Array, int]
@@ -451,7 +451,8 @@ class MaskedState(NamedTuple):
   """Maintains inner transform state for masked transformations.
   
   Attributes:
-    inner_state (:class:`optax.OptState`): The state of the inner `GradientTransformation`.
+    inner_state (:class:`optax.OptState`): The state of the inner
+      `GradientTransformation`.
   """
   inner_state: base.OptState
 
@@ -570,7 +571,8 @@ class MaybeUpdateState(NamedTuple):
   """Maintains inner transform state and adds a step counter.
   
   Attributes:
-    inner_state (:class:`optax.OptState`): The state of the inner `GradientTransformation`.
+    inner_state (:class:`optax.OptState`): The state of the inner
+      `GradientTransformation`.
     step (``Union[jax.Array, int]``): The current step counter.
   """
   inner_state: base.OptState