diff --git a/docs/cookbook/nemotron_qa.md b/docs/cookbook/nemotron_qa.md
index 126ff11..2cb76c0 100644
--- a/docs/cookbook/nemotron_qa.md
+++ b/docs/cookbook/nemotron_qa.md
@@ -1,14 +1,14 @@
 ---
 categories:
-- Preference Data
-description: Learn to implement Nvidia's Preference Data Pipeline with Dria, using
-  synthetic data generation and reward modeling techniques.
+- Synthetic Data
+description: Learn to implement Nvidia's Preference Data Pipeline using Dria for synthetic
+  data generation with Llama 3.1.
 tags:
 - Nvidia
-- Dria
-- Data Pipeline
-- Machine Learning
 - Synthetic Data
+- Dria
+- AI Pipeline
+- Llama 3.1
 ---
 
 # Implementing Nvidia's Preference Data Pipeline with Dria
diff --git a/docs/how-to/data_enrichment.md b/docs/how-to/data_enrichment.md
index 753a3a6..4dc08db 100644
--- a/docs/how-to/data_enrichment.md
+++ b/docs/how-to/data_enrichment.md
@@ -28,11 +28,11 @@ tags:
 Here's a complete example showing how to analyze the extract summary of a text using Dria:
 
 ```python
-# Define the schema for translated content
+# Define the schema for summarized content
 class SummarizedContent(BaseModel):
     summary: str
 
-# Create a prompt with the translation instruction
+# Create a prompt with the summary instruction
 prompter = Prompt(
    "Summarize the following text in a single concise paragraph:\n\n{{text}}",
    schema=SummarizedContent
@@ -126,7 +126,7 @@ async def enrich():
       sentiment: str
       keywords: str
 
-   # Create a prompt with the translation instruction
+   # Create a prompt with the analysis instruction
    prompter = Prompt(
       "Identify the sentiment (positive, negative, or neutral) of the following text and extract keywords:\n\n{{generation}}",
       schema=AnalyzedText