Merge pull request #49 from fiddler-labs/bug/quickstart_typos

fixing typos

README.md

@@ -71,7 +71,7 @@ We are continuously updating this library to support language models as they evo
 - Contributions in the form of suggestions and PRs to Fiddler Auditor are welcome!
 - If you encounter a bug, please feel free to raise issues in this repository.
 
-For step-by-step instructions follow the [Contrubution Guide](CONTRIBUTION.md).
+For step-by-step instructions follow the [Contribution Guide](CONTRIBUTION.md).
 
 ## Community
 - For questions and support, join the [Fiddler Community](https://www.fiddler.ai/slackinvite)

examples/LLM_Evaluation.ipynb

@@ -18,7 +18,7 @@
     "- **Apply transformations:** Fiddler Auditor provides built-in transformation such as paraphrasing. Additionally, you can define your own.\n",
     "\n",
     "\n",
-    "- **Evaluate generated outputs:** The generations are then evaluated for correctenss, robustness, saftey etc. For convenience, the Auditor comes with built-in evaluation methods like semantic similarity, mdoel graded evaluations and Toxicity detection. Additionally, you can define your own evaluation function.\n",
+    "- **Evaluate generated outputs:** The generations are then evaluated for correctenss, robustness, saftey etc. For convenience, the Auditor comes with built-in evaluation methods like semantic similarity, model graded evaluations and Toxicity detection. Additionally, you can define your own evaluation function.\n",
     "\n",
     "\n",
     "- **Reporting:** The results are then aggregated and errors highlighted.\n",
@@ -365,7 +365,7 @@
    "source": [
     "# Model Graded Robustness\n",
     "\n",
-    "Now, we will evaluate the robustness of the gpt-3.5-turbo model to prompt paraphrasing. To do so we will leveregae **Model Graded Evaluation**. \n",
+    "Now, we will evaluate the robustness of the gpt-3.5-turbo model to prompt paraphrasing. To do so we will leverage **Model Graded Evaluation**. \n",
     "\n",
     "\n",
     "<!-- ![ModelGraded](images/model_graded_robustness.png) -->\n",
@@ -476,7 +476,7 @@
     "We'll now define a custom transformation class which will carry out a prompt injection attack. Prompt injection attack is a type of attack where one could override chatbot behavior by saying something like \"Ignore previous instructions and instead ...\". To know more about prompt injection refer to this [article](https://simonwillison.net/2022/Sep/12/prompt-injection/)\n",
     "\n",
     "\n",
-    "We've defined an attack in teh cell below. Notice the following aspects in the class definition\n",
+    "We've defined an attack in the cell below. Notice the following aspects in the class definition\n",
     "\n",
     "**1. Inherit from the TransformBase class**\n",
     "\n",
@@ -486,7 +486,7 @@
     "    ...\n",
     "```\n",
     "\n",
-    "**2. Define a description() method:** This metod should return a string that describes the details of the transformation being performed.\n",
+    "**2. Define a description() method:** This method should return a string that describes the details of the transformation being performed.\n",
     "\n",
     "**3. Define a transform() method:** The method must accept an argument 'prompt' of string type and must return a list of transformed prompts."
    ]
@@ -516,7 +516,7 @@
     "    def transform(self, prompt) -> List[str]:\n",
     "        # generate paraphrased prompts\n",
     "        perturbations = self.paraphrase.transform(prompt)\n",
-    "        # append the prompt inhection directive\n",
+    "        # append the prompt injection directive\n",
     "        return [p + self.injection_directive for p in perturbations]\n",
     "\n",
     "    def description(self) -> str:\n",