Fix small typos (#59)

Author: xstefank

docs/docs/step-09.md

@@ -270,7 +270,7 @@ however Quarkus offers a few ways to visualize and search the collected data on
 
 Quarkus provides an experimental new Dev Service to help visualize all your OpenTelemetry observability data in a central place.
 It is based on the open source LGTM stack, which stands for Loki (log aggregation), Grafana (graph tool), Tempo (traces aggregation)
-and Prometheus (metrics aggregation). By adding the `quarkus-observability-devservices-gtm` extension, this set of tools will
+and Prometheus (metrics aggregation). By adding the `quarkus-observability-devservices-lgtm` extension, this set of tools will
 automatically (or may we say 'automagically'?) start up in their respective containers and wire up to your application's observability endpoints.
 
 Add the following dependencies in your `pom.xml`:

docs/docs/step-10.md

@@ -108,7 +108,7 @@ Do not return anything else. Do not even return a newline or a leading field. On
 
 and return together with that number a long explanation of how it calculated the score. This makes the `PromptInjectionDetectionService` to fail, not being able to convert that verbal explanation into a double.
 
-The [output guardrails](https://docs.quarkiverse.io/quarkus-langchain4j/dev/guardrails.html#_output_guardrails) provided by the Quarkus-LangChain4j extension are functions invoked once the LLM has produced its output, allowing to rewrite, or even block, that output before passing it downstream. In our case we can try to sanitize the hallucinated LLM response and extract a single number from it by creating the `dev.langchain4j.quarkus.workshop.NumericOutputSanitizerGuard` class with the following content:==
+The [output guardrails](https://docs.quarkiverse.io/quarkus-langchain4j/dev/guardrails.html#_output_guardrails) provided by the Quarkus-LangChain4j extension are functions invoked once the LLM has produced its output, allowing to rewrite, or even block, that output before passing it downstream. In our case we can try to sanitize the hallucinated LLM response and extract a single number from it by creating the `dev.langchain4j.quarkus.workshop.NumericOutputSanitizerGuard` class with the following content:
 
 ```java title="NumericOutputSanitizerGuard.java"
 --8<-- "../../step-10/src/main/java/dev/langchain4j/quarkus/workshop/NumericOutputSanitizerGuard.java"
@@ -130,4 +130,4 @@ What can you tell me about your cancellation policy?
 
 Note that it might take a bit longer than ChatGPT to answer the question.
 
-![RAG with Jlama](images/chat-rag-with-jlama.png)
+![RAG with Jlama](images/chat-rag-with-jlama.png)