diff --git a/tensorflow_model_optimization/g3doc/guide/pruning/comprehensive_guide.ipynb b/tensorflow_model_optimization/g3doc/guide/pruning/comprehensive_guide.ipynb
index d674f47f0..423b7e767 100644
--- a/tensorflow_model_optimization/g3doc/guide/pruning/comprehensive_guide.ipynb
+++ b/tensorflow_model_optimization/g3doc/guide/pruning/comprehensive_guide.ipynb
@@ -387,7 +387,7 @@
         "1. Prune a custom Keras layer\n",
         "2. Modify parts of a built-in Keras layer to prune.\n",
         "\n",
-        "For an example, the API defaults to only pruning the kernel of the\n",
+        "For example, the API defaults to only pruning the kernel of the\n",
         "`Dense` layer. The example below prunes the bias also.\n"
       ]
     },
@@ -587,7 +587,7 @@
         "\n",
         "* Have a learning rate that's not too high or too low when the model is pruning. Consider the [pruning schedule](https://www.tensorflow.org/model_optimization/api_docs/python/tfmot/sparsity/keras/PruningSchedule) to be a hyperparameter.\n",
         "\n",
-        "* As a quick test, try experimenting with pruning a model to the final sparsity at the begining of training by setting `begin_step` to 0 with a `tfmot.sparsity.keras.ConstantSparsity` schedule. You might get lucky with good results.\n",
+        "* As a quick test, try experimenting with pruning a model to the final sparsity at the beginning of training by setting `begin_step` to 0 with a `tfmot.sparsity.keras.ConstantSparsity` schedule. You might get lucky with good results.\n",
         "\n",
         "* Do not prune very frequently to give the model time to recover. The [pruning schedule](https://www.tensorflow.org/model_optimization/api_docs/python/tfmot/sparsity/keras/PruningSchedule) provides a decent default frequency.\n",
         "\n",
@@ -723,7 +723,7 @@
         "id": "yqk0jI49c1mw"
       },
       "source": [
-        "Once different backends [enable pruning to improve latency]((https://github.com/tensorflow/model-optimization/issues/173)), using block sparsity can improve latency for certain hardware.\n",
+        "Once different backends [enable pruning to improve latency](https://www.tensorflow.org/model_optimization/guide/pruning), using block sparsity can improve latency for certain hardware.\n",
         "\n",
         "Increasing the block size will decrease the peak sparsity that's achievable for a target model accuracy. Despite this, latency can still improve.\n",
         "\n",