technical edits per Annie

Author: kknoxrht

modules/ROOT/pages/index.adoc

@@ -41,7 +41,7 @@ This course will be experienced from the perspective of a *development team* at
 == Prerequisites
 
 
-* Basic knowledge of the Large Language mMdel (LLM) terminology.
+* Basic knowledge of the Large Language Model (LLM) terminology.
 * Interest in the simplest way to interact with AI Models.
 * Hardware with 4vCPUs, 8GB of memory, and 20-60GB of disk space for a good experience.
 * MacOS, Windows, Linux Operating System with a GUI or Desktop enabled.

modules/chapter1/pages/index.adoc

@@ -8,7 +8,7 @@ What tools, services, and solutions are available today that will allow my produ
 
  * What offers the least friction in development? 
  * How can we guide this with accelerators that don't require months of build time? 
- * How can we allow our product teams to focus on our company's core differentiators while combining the power of Large Language Models (LLMs) to innovate on behalf of our customers?
+ * How can we allow our product teams to focus on our company's core differentiators while combining the power of large language models (LLMs) to innovate on behalf of our customers?
 
 The GizmoGobble's leadership team decided that when it comes to LLM development, they should exhaust the path of least custom development before moving to more complicated and costly approaches. 
 

modules/chapter1/pages/section1.adoc

@@ -19,7 +19,7 @@ They turned to *Podman AI Lab*, an open source extension for local AI developmen
 Excited by this new opportunity, GizmoGobble's developers began exploring the offerings of Podman AI Lab. They were particularly intrigued by the Recipes Catalog, which provided detailed explanations and sample applications for various core AI use cases like ChatBots, Code Generators, and Text Summarizers. By experimenting with multiple AI models, they hoped to find the optimal one for their unique ChatBot application.
 
 
-Next, they discovered the Model Catalog, a curated list of open source Generative AI Large Language Models (LLMs) provided by Podman AI Lab. Once downloaded, these models could be used for various AI applications, model services, and playgrounds,  enabling model evaluation and experimentation.
+Next, they discovered the Model Catalog, a curated list of open source Generative AI large language models (LLMs) provided by Podman AI Lab. Once downloaded, these models could be used for various AI applications, model services, and playgrounds,  enabling model evaluation and experimentation.
 
 GG's team was eager to incorporate these advanced technologies into GizmoGenie.
 With a newfound sense of determination and the intuitive user interface of Podman AI Lab, it was easy for them to explore and learn about AI technology.

modules/chapter4/pages/section1.adoc

@@ -4,7 +4,7 @@
 
 RAG (Retrieval-Augmented Generation) is an AI framework that combines the strengths of traditional information retrieval systems (such as databases) with the capabilities of generative large language models (LLMs).
 
-Retrieval-Augmented Generation (RAG) is the process of optimizing the output of a large language model, so it references an authoritative knowledge base outside of its training data sources before generating a response. Large Language Models (LLMs) are trained on vast volumes of data and use billions of parameters to generate original output for tasks like answering questions, translating languages, and completing sentences. 
+Retrieval-Augmented Generation (RAG) is the process of optimizing the output of a large language model, so it references an authoritative knowledge base outside of its training data sources before generating a response. Large language models (LLMs) are trained on vast volumes of data and use billions of parameters to generate original output for tasks like answering questions, translating languages, and completing sentences. 
 
 RAG extends the already powerful capabilities of LLMs to specific domains or an organization's internal knowledge base, all without the need to retrain the model. It is a cost-effective approach to improving LLM output so it remains relevant, accurate, and useful in various contexts.
 
@@ -17,7 +17,7 @@ Known challenges of LLMs include:
  * Creating a response from non-authoritative sources.
  * Creating inaccurate responses due to terminology confusion, wherein different training sources use the same terminology to talk about different things.
 
-You can think of the Large Language Model as an over-enthusiastic new employee who refuses to stay informed with current events but will always answer every question with absolute confidence. Unfortunately, such an attitude can negatively impact user trust and is not something you want your ChatBots to emulate!
+You can think of the large language model as an over-enthusiastic new employee who refuses to stay informed with current events but will always answer every question with absolute confidence. Unfortunately, such an attitude can negatively impact user trust and is not something you want your ChatBots to emulate!
 
 RAG is one approach to solving some of these challenges. It redirects the LLM to retrieve relevant information from authoritative, pre-determined knowledge sources. Organizations have greater control over the generated text output, and users gain insights into how the LLM generates the response.
 
@@ -52,7 +52,7 @@ The following lab is an excellent example of the steps the GizmoGobble developer
 
 Podman AI Lab is an excellent place to evaluate and test models, but you'll eventually want to see how this will actually be deployed in your enterprise. For that, we can use OpenShift and OpenShift AI along with the Elasticsearch vector database to create a Retrieval Augmented Generation (RAG) integrated ChatBot.
 
-==== lab overview and requirements
+==== Lab Overview and Requirements
 
 This above article will direct you to website which will walk you through how to go from a ChatBot recipe in the Podman AI Lab extension to a RAG ChatBot deployed on OpenShift and OpenShift AI.
 

modules/chapter4/pages/section2.adoc

@@ -19,12 +19,12 @@ The LAB method consists of 3 components:
 
  . *Taxonomy-driven data curation.* Taxonomy is a set of diverse training data curated by humans as examples of new knowledge and skills for the model.
  . *Large-scale synthetic data generation.* The model is then used to generate new examples based on the seed training data. Recognizing that synthetic data can vary in quality, the LAB method adds an automated step to refine the example answers, making sure they’re grounded and safe.
- . *Iterative, large-scale alignment tuning.* Finally, the model is retrained based on the set of synthetic data. The LAB method includes 2 tuning phases: knowledge tuning, followed by skill tuning.
+ . *Iterative, large-scale alignment tuning.* Finally, the model is retrained based on the set of synthetic data. The LAB method includes two tuning phases: knowledge tuning, followed by skill tuning.
 
 
 
 === Alignment tuning
-After pretraining, LLMs undergo alignment tuning to make the model’s answers as accurate and useful as possible. The 1st step in alignment tuning is typically instruction tuning, in which a model is trained directly on specific tasks of interest. Next is preference tuning, which can include *reinforcement learning from human feedback* (RLHF). In this step, humans test the model and rate its output, noting if the model’s answers are preferred or unpreferred. An RLHF process may include multiple rounds of feedback and refinement to optimize a model.
+After pretraining, LLMs undergo alignment tuning to make the model’s answers as accurate and useful as possible. The first step in alignment tuning is typically instruction tuning, in which a model is trained directly on specific tasks of interest. Next is preference tuning, which can include *reinforcement learning from human feedback* (RLHF). In this step, humans test the model and rate its output, noting if the model’s answers are preferred or unpreferred. An RLHF process may include multiple rounds of feedback and refinement to optimize a model.
 
 Researchers have found that the amount of feedback at this alignment tuning stage can be much smaller than the initial set of training data―tens of thousands of human annotations, compared to the trillions of tokens of data required for pretraining―and still unlock latent capabilities of the model.
 
@@ -55,7 +55,9 @@ Take a look at "lab-enhanced" models on the https://huggingface.co/instructlab[I
 
 ---
 
-===  Optional Lab: https://github.com/RedHatQuickCourses/instructlab-insurance-lab/tree/main[InstructLab Unleashed] --*COMING SOON*
+// ===  Optional Lab: https://github.com/RedHatQuickCourses/instructlab-insurance-lab/tree/main
+
+=== InstructLab Unleashed --*COMING SOON*
 
 
 InstructLab Unleashed: Training the Merlinite Model to Answer Questions for Parasol Insurance!