Acrolinx updates

Author: theresa-i

learn-pr/wwl-data-ai/multistage-reasoning-azure-databricks/includes/02-multi-stage-reasoning-concepts.md

@@ -28,7 +28,7 @@ The core idea of multi-stage reasoning systems is that by dividing a problem int
 
 ## Design a multi-stage reasoning system
 
-When you think your solution will benefit from implementing a multi-stage reasoning system, you need to map your requirements to the necessary tools and components.
+When you think your solution benefits from implementing a multi-stage reasoning system, you need to map your requirements to the necessary tools and components.
 
 To design a multi-stage reasoning system, you break down your problem into multiple steps, or tasks. As shown in the sentiment analysis example, you can reuse the output of a previous task as input for a new task, by **chaining** LLM interactions.
 
@@ -47,8 +47,8 @@ By breaking up your problem and designing a multi-stage reasoning system, you cr
 After your initial design is complete, you can choose one of the popular Python libraries to build your multi-stage reasoning system. Understanding multiple frameworks is valuable because:
 
 - **Different approaches**: Each framework takes a unique approach to solving multi-stage reasoning challenges
-- **Project requirements**: Your specific use case may favor one framework's strengths over others
-- **Team expertise**: Your organization may already have experience with a particular framework
+- **Project requirements**: Your specific use case might favor one framework's strengths over others
+- **Team expertise**: Your organization might already have experience with a particular framework
 - **Integration needs**: Some frameworks integrate better with existing systems or data sources
 
 Azure Databricks provides access to several frameworks designed for multi-stage reasoning:

learn-pr/wwl-data-ai/multistage-reasoning-azure-databricks/includes/03-langchain.md

@@ -1,4 +1,4 @@
-LangChain is an open-source Python library designed to simplify building applications with Large Language Models (LLMs). LangChain provides tools to chain together multiple LLM interactions in a structured workflow. Use this framework when you want to build complex workflows or need access to many pre-built tools and integrations.
+LangChain is an open-source Python library designed to simplify building applications with Large Language Models (LLMs). LangChain provides tools to chain together multiple LLM interactions in a structured workflow. Use this framework when you want to build complex workflows or need access to many prebuilt tools and integrations.
 
 LangChain abstracts the complexity of connecting different steps in your reasoning pipeline, allowing you to focus on designing the logic rather than managing the technical details of how components communicate.
 
@@ -29,7 +29,7 @@ Well-designed prompts ensure each stage of your reasoning process produces the r
 
 ### Find context with a retriever
 
-A **retriever** finds and returns relevant information from external sources like documents or databases. In multi-stage reasoning, retrievers typically gather context that subsequent language model operations will use to generate responses.
+A **retriever** finds and returns relevant information from external sources like documents or databases. In multi-stage reasoning, retrievers typically gather context that subsequent language model operations use to generate responses.
 
 Retrievers often work with **vector indices** and **embeddings** to find semantically similar content. Embeddings convert text into numerical vectors that capture meaning, while vector indices enable fast similarity searches across large datasets.
 
@@ -50,7 +50,7 @@ To use LangChain in Azure Databricks, you need to ensure the library is installe
 %pip install langchain
 ```
 
-Once installed, you can use LangChain to build multi-stage reasoning systems. Here's how you can structure a simple multi-stage workflow:
+Once installed, you can use LangChain to build multi-stage reasoning systems. Here's how you can structure a multi-stage workflow:
 
 ### Conceptual Example: Question Answering with Document Search
 

learn-pr/wwl-data-ai/multistage-reasoning-azure-databricks/includes/06-dspy-framework.md

@@ -1,6 +1,6 @@
 **DSPy** is a framework that helps you build programs with Large Language Models (LLMs) by focusing on what you want the model to do rather than how to prompt it. Instead of manually writing and adjusting prompts, DSPy automatically optimizes prompts and system performance for you. Use this framework when you want systematic optimization of your LLM interactions rather than manual prompt engineering.
 
-DSPy treats LLM interactions as programmable modules that can be composed, optimized, and reused. When building multi-stage reasoning systems, DSPy lets you declare the behavior you need at each stage (using signatures) and automatically finds the best way to prompt the LLM to achieve that behavior.
+DSPy treats LLM interactions as programmable modules that can be composed, optimized, and reused. When you build multi-stage reasoning systems, DSPy lets you declare the behavior you need at each stage (using signatures) and automatically finds the best way to prompt the LLM to achieve that behavior.
 
 ## Explore the main components of DSPy