diff --git a/tutorials/00_NeMo_Primer.ipynb b/tutorials/00_NeMo_Primer.ipynb
index c21696702a39..18b8652fa10a 100644
--- a/tutorials/00_NeMo_Primer.ipynb
+++ b/tutorials/00_NeMo_Primer.ipynb
@@ -588,7 +588,7 @@
         "id": "U7Eezf_sAVS0"
       },
       "source": [
-        "You might wonder why we didnt explicitly set `citrinet.cfg.optim = cfg.optim`. \n",
+        "You might wonder why we didn't explicitly set `citrinet.cfg.optim = cfg.optim`. \n",
         "\n",
         "This is because the `setup_optimization()` method does it for you! You can still update the config manually."
       ]
diff --git a/tutorials/asr/ASR_Confidence_Estimation.ipynb b/tutorials/asr/ASR_Confidence_Estimation.ipynb
index e177a5132b26..06bb75f8f237 100644
--- a/tutorials/asr/ASR_Confidence_Estimation.ipynb
+++ b/tutorials/asr/ASR_Confidence_Estimation.ipynb
@@ -284,7 +284,7 @@
     "            eps_padded_hyp, labels, padded_labels, fill_confidence_deletions(confidence_scores, labels)\n",
     "        ):\n",
     "            word_len = len(word)\n",
-    "            # shield angle brakets for <eps>\n",
+    "            # shield angle brackets for <eps>\n",
     "            if html and word == \"<eps>\":\n",
     "                word = \"&lt;eps&gt;\"\n",
     "            if current_line_len + word_len + 1 <= terminal_width:\n",
@@ -307,7 +307,7 @@
     "        current_word_line = \"\"\n",
     "        for word, score in zip(transcript_list, confidence_scores):\n",
     "            word_len = len(word)\n",
-    "            # shield angle brakets for <eps>\n",
+    "            # shield angle brackets for <eps>\n",
     "            if html and word == \"<eps>\":\n",
     "                word = \"&lt;eps&gt;\"\n",
     "            if current_line_len + word_len + 1 <= terminal_width:\n",
diff --git a/tutorials/asr/ASR_Context_Biasing.ipynb b/tutorials/asr/ASR_Context_Biasing.ipynb
index 75385234ce29..ec8c0c1b78c6 100644
--- a/tutorials/asr/ASR_Context_Biasing.ipynb
+++ b/tutorials/asr/ASR_Context_Biasing.ipynb
@@ -361,7 +361,7 @@
    "source": [
     "## Create a context-biasing list\n",
     "\n",
-    "Now, we need to select the words, recognition of wich we want to improve by CTC-WS context-biasing.\n",
+    "Now, we need to select the words, recognition of which we want to improve by CTC-WS context-biasing.\n",
     "Usually, we select only nontrivial words with the lowest recognition accuracy.\n",
     "Such words should have a character length >= 3 because short words in a context-biasing list may produce high false-positive recognition.\n",
     "In this toy example, we will select all the words that look like names with a recognition accuracy less than 1.0.\n",
diff --git a/tutorials/asr/Speech_Commands.ipynb b/tutorials/asr/Speech_Commands.ipynb
index f0671763b984..e50e8d1f283e 100644
--- a/tutorials/asr/Speech_Commands.ipynb
+++ b/tutorials/asr/Speech_Commands.ipynb
@@ -1431,10 +1431,10 @@
                 "# Lets change the scheduler\n",
                 "optim_sched_cfg.sched.name = \"CosineAnnealing\"\n",
                 "\n",
-                "# \"power\" isnt applicable to CosineAnnealing so let's remove it\n",
+                "# \"power\" isn't applicable to CosineAnnealing so let's remove it\n",
                 "optim_sched_cfg.sched.pop('power')\n",
                 "\n",
-                "# \"hold_ratio\" isnt applicable to CosineAnnealing, so let's remove it\n",
+                "# \"hold_ratio\" isn't applicable to CosineAnnealing, so let's remove it\n",
                 "optim_sched_cfg.sched.pop('hold_ratio')\n",
                 "\n",
                 "# Set \"min_lr\" to lower value\n",
diff --git a/tutorials/nlp/Joint_Intent_and_Slot_Classification.ipynb b/tutorials/nlp/Joint_Intent_and_Slot_Classification.ipynb
index a0ff0faf511b..b21fdfe36020 100644
--- a/tutorials/nlp/Joint_Intent_and_Slot_Classification.ipynb
+++ b/tutorials/nlp/Joint_Intent_and_Slot_Classification.ipynb
@@ -749,7 +749,7 @@
             "source": [
                 "### Optimizing Threshold\n",
                 "\n",
-                "As mentioned above, when classifiying a given query such as `show all flights and fares from denver to san francisco`, our model checks whether each individual intent would be suitable. Before assigning the final labels for a query, the model assigns a probability an intent matches the query. For example, if our `dict.intents.csv` had 5 different intents, then the model could output for a given query \\[0.52, 0.38, 0.21, 0.67. 0.80\\] where each value represents the probability that query matches that particular intent. \n",
+                "As mentioned above, when classifying a given query such as `show all flights and fares from denver to san francisco`, our model checks whether each individual intent would be suitable. Before assigning the final labels for a query, the model assigns a probability an intent matches the query. For example, if our `dict.intents.csv` had 5 different intents, then the model could output for a given query \\[0.52, 0.38, 0.21, 0.67. 0.80\\] where each value represents the probability that query matches that particular intent. \n",
                 "\n",
                 "We need to use these probabilities to generate final label predictions of 0 or 1 for each label. While we can use 0.5 as the probability threshold, it is usually the case that there is a better threshold to use depending on the metric we want to optimize. For this tutorial, we will be finding the threshold that gives us the best micro-F1 score on the validation set. After running the `optimize_threshold` method, the threshold attribute for our model will be updated."
             ]