AWS solutions documentation change
Summary
Updated typographical formatting by replacing straight apostrophes with curly apostrophes in possessive forms (e.g., 'user's' to 'user’s')
Security assessment
Changes are purely typographical/formatting improvements with no security context. No security vulnerabilities, mitigations, or security features are mentioned in the diff.
Diff
diff --git a/solutions/latest/generative-ai-application-builder-on-aws/configuring-your-prompts.md index e32d94f5e..ddaf94240 100644 --- a/solutions/latest/generative-ai-application-builder-on-aws/configuring-your-prompts.md +++ b/solutions/latest/generative-ai-application-builder-on-aws/configuring-your-prompts.md @@ -11 +11 @@ This section controls the overall experience and behavior of the AI prompt. - * **Max prompt template length** : This setting determines the maximum length (in characters) of the prompt template. A higher value allows for more context to be provided to the AI model, potentially leading to more accurate responses. However, excessively long prompts may also introduce noise and negatively impact performance. For Amazon Bedrock models, the default values for max prompt template length (in characters) is calculated using the underlying model token limits. If you edit and change a model name within Bedrock, 'Reset to default' button is highlighted and can be used to adopt the newly selected model's defaults. For Amazon SageMaker models, reasonable default values are provided, but it is recommended that you check your underlying model and choose these max prompt template length and input text lengths accordingly. Refer to the Tips on managing model token limits section for more information. + * **Max prompt template length** : This setting determines the maximum length (in characters) of the prompt template. A higher value allows for more context to be provided to the AI model, potentially leading to more accurate responses. However, excessively long prompts may also introduce noise and negatively impact performance. For Amazon Bedrock models, the default values for max prompt template length (in characters) is calculated using the underlying model token limits. If you edit and change a model name within Bedrock, 'Reset to default' button is highlighted and can be used to adopt the newly selected model’s defaults. For Amazon SageMaker models, reasonable default values are provided, but it is recommended that you check your underlying model and choose these max prompt template length and input text lengths accordingly. Refer to the Tips on managing model token limits section for more information. @@ -13 +13 @@ This section controls the overall experience and behavior of the AI prompt. - * **Max input text length** : This setting limits the maximum length (in characters) of the user's input text. Longer inputs may contain irrelevant information, increasing the risk of obtaining irrelevant or inaccurate responses from the AI model. + * **Max input text length** : This setting limits the maximum length (in characters) of the user’s input text. Longer inputs may contain irrelevant information, increasing the risk of obtaining irrelevant or inaccurate responses from the AI model. @@ -22 +22 @@ This section controls the overall experience and behavior of the AI prompt. -This section allows you to define the actual prompt template that will be used by the AI model. The prompt template typically follows a structure that includes placeholders for various components, such as the user's input, reference passages, and chat history. +This section allows you to define the actual prompt template that will be used by the AI model. The prompt template typically follows a structure that includes placeholders for various components, such as the user’s input, reference passages, and chat history. @@ -26 +26 @@ This section allows you to define the actual prompt template that will be used b - * `{input}`: This placeholder is mandatory and will be substituted with the user's input or query. + * `{input}`: This placeholder is mandatory and will be substituted with the user’s input or query. @@ -32 +32 @@ This section allows you to define the actual prompt template that will be used b - * **Rephrase Question?** : This option (available for RAG deployments only) determines whether the user's original input query should be rephrased or disambiguated before being passed to the AI model. Rephrasing the query can sometimes help the model better understand the user's intent, potentially leading to more accurate responses. + * **Rephrase Question?** : This option (available for RAG deployments only) determines whether the user’s original input query should be rephrased or disambiguated before being passed to the AI model. Rephrasing the query can sometimes help the model better understand the user’s intent, potentially leading to more accurate responses. @@ -37 +37 @@ This section allows you to define the actual prompt template that will be used b -When configuring the prompt template and experience, it's essential to strike a balance between providing sufficient context and instructions to the AI model while avoiding excessively long or irrelevant information that may introduce noise or performance issues. +When configuring the prompt template and experience, it’s essential to strike a balance between providing sufficient context and instructions to the AI model while avoiding excessively long or irrelevant information that may introduce noise or performance issues. @@ -60 +60 @@ This section allows you to configure the behavior and template for disambiguatin - * **Disambiguation Prompt Template** : This is the prompt template used for disambiguating user inputs when connected to a knowledge base. The output generated from this prompt will be used as the query sent to the knowledge base. Disabling disambiguation would result in the user's raw query being sent to the knowledge base unchanged. + * **Disambiguation Prompt Template** : This is the prompt template used for disambiguating user inputs when connected to a knowledge base. The output generated from this prompt will be used as the query sent to the knowledge base. Disabling disambiguation would result in the user’s raw query being sent to the knowledge base unchanged.