diff --git a/minitorch/tensor_ops.py b/minitorch/tensor_ops.py
index 2eb9a5d0..9cc82e25 100644
--- a/minitorch/tensor_ops.py
+++ b/minitorch/tensor_ops.py
@@ -6,6 +6,8 @@
     shape_broadcast,
     MAX_DIMS,
 )
+from minitorch import Tensor
+from typing import Callable
 
 
 def tensor_map(fn):
@@ -44,7 +46,7 @@ def _map(out, out_shape, out_strides, in_storage, in_shape, in_strides):
     return _map
 
 
-def map(fn):
+def map(fn: Callable[[float], float]) -> Callable[[Tensor], Tensor]:
     """
     Higher-order tensor map function ::
 
@@ -71,7 +73,8 @@ def map(fn):
                should broadcast with `a`
 
     Returns:
-        :class:`TensorData` : new tensor data
+        function: A function that takes a tensor, applies `fn` to each cell,
+        and returns the resulting tensor.
     """
 
     f = tensor_map(fn)
@@ -134,7 +137,7 @@ def _zip(
     return _zip
 
 
-def zip(fn):
+def zip(fn: Callable[[float, float], float]) -> Callable[[Tensor, Tensor], Tensor]:
     """
     Higher-order tensor zip function ::
 
@@ -160,7 +163,8 @@ def zip(fn):
         b (:class:`TensorData`): tensor to zip over
 
     Returns:
-        :class:`TensorData` : new tensor data
+        function: A function that takes two tensors, applies `fn` to each pair
+        of cells, and returns the resulting tensor.
     """
 
     f = tensor_zip(fn)
@@ -204,7 +208,7 @@ def _reduce(out, out_shape, out_strides, a_storage, a_shape, a_strides, reduce_d
     return _reduce
 
 
-def reduce(fn, start=0.0):
+def reduce(fn: Callable[[float, float], float], start: float = 0.0) -> Callable[[Tensor, float], Tensor]:
     """
     Higher-order tensor reduce function. ::
 
@@ -225,7 +229,9 @@ def reduce(fn, start=0.0):
         dim (int): int of dim to reduce
 
     Returns:
-        :class:`TensorData` : new tensor
+        function: A function that takes a tensor, applies `fn` to all cells
+        along a particular axis (or all cells if `dim` is not specified), and
+        returns the resulting tensor.
     """
     f = tensor_reduce(fn)